Data Integration Guide. salesforce Wave Analytics

Create Datasets with a Dataflow

A dataset is a collection of related data that is stored in a denormalized, yet highly compressed form. For each platform license, your organization can have a maximum of 250 million rows of data stored for all registered datasets combined.

You can use a dataflow to create one or more datasets based on data from Salesforce objects or existing datasets.

Transformations for Wave Dataflows

A transformation refers to the manipulation of data. You can add transformations to a dataflow to extract data from Salesforce objects or datasets, transform datasets that contain Salesforce or external data, and register datasets.

Create a Dataset with the Dataset Builder

The data in the dataset refreshes each time the dataflow runs. You can also edit the dataflow definition file to add transformations that manipulate the dataset.

Optimize Your Dataflows with Replication

Replication decouples the extract of Salesforce data from your dataflows, letting you run it on a separate schedule. By scheduling these extracts ahead of time, your dataflow has less to do and runs faster. To lighten the load even more, replication extracts Salesforce data incrementally by default, meaning that only data that’s changed gets extracted.

Wave Connector for Excel Data

The Salesforce Wave Connector makes it easy to import data from Microsoft Excel 2013 to Wave Analytics.

Create a Dataset with External Data

You can either upload external data through the user interface or through the External Data API to create a dataset. When you upload an external data file (in .csv, .gz, or .zip format), you can also provide a metadata file. A metadata file contains metadata attributes that describe the structure of the data in the external data file. If you upload a .csv from the user interface, Wave Analytics automatically generates the metadata file, which you can preview and change. If you do not provide a metadata file, Wave Analytics imports all external data file columns as dimensions.

Edit a Dataset

You can edit a dataset to change the dataset name, app, security predicate, or extended metadata (XMD) file associated with the dataset. For datasets created from an external data file, you can also upload a new external data file or metadata file to update the data or metadata.

Analytics Data Integration

Delete a Dataset

Delete unnecessary datasets from your My Private App or in shared apps on which you have at least Editor access. Removing datasets reduces clutter and helps you avoid reaching your org's limit for rows across registered datasets.

Create Datasets with Recipes in Data Prep (Beta)

Data prep gives you user interface tools for managing, monitoring, and creating datasets. Using recipes, you can combine data from multiple datasets or objects, remove fields, and transform field values right within the data manager.

Row-Level Security for Datasets

If a Wave Analytics user has access to a dataset, they have access to all records in the dataset, by default. However, you can implement row-level security on a dataset to restrict access to records. Some records might contain sensitive data that shouldn’t be accessible by everyone.

Predicate Expression Syntax for Datasets

You must use valid syntax when defining the predicate expression.

Datasets

Wave Analytics applies one of the following types to each dataset field:

Date

A date can be represented as a day, month, year, and, optionally, time. You can group, filter, and perform math on dates.

Dimension

A dimension is a qualitative value, like region, product name, and model number. Dimensions are handy for grouping and filtering your data. Unlike measures, you can’t perform math on dimensions. To increase query performance, Wave Analytics indexes all dimension fields in datasets.

Measure

A measure is a quantitative value, like revenue and exchange rate. You can do math on measures, such as calculating the total revenue and minimum exchange rate.

For each dataset that you create, you can apply row-level security to restrict access to records in the dataset.

Attention: Before you create a dataset, verify that the source data contains at least one value in each column. Columns with all null values won't be created in datasets and can't be referenced in dataflows, lenses, or dashboards. Consider providing a default value for null values, like "n/a" or "empty."

IN THIS SECTION:

Numeric-Value Handling in Datasets

Wave Analytics internally stores numeric values in datasets as long values. For example, it stores the number “3,200.99” with a scale of “2” as “320099”. The user interface converts the stored value back to decimal notation to display the number as “3200.99.”

Date Handling in Datasets

When Wave Analytics loads dates into a dataset, it breaks up each date into multiple fields, such as day, week, month, quarter, and year, based on the calendar year. For example, if you extract dates from a

CreateDate

field, Wave Analytics generates date fields such as

CreateDate_Day

and

CreateDate_Week

. If your fiscal year differs from the calendar year, you can enable Wave

Analytics to generate fiscal date fields as well.

Analytics Data Integration

Dataflow JSON

You can use the dataflow to create one or more datasets based on data from Salesforce objects or existing datasets. A dataflow is a set of instructions that specifies what data to extract from Salesforce objects or datasets, how to transform the datasets, and which datasets to make available for querying. With a dataflow, you can manipulate the extracted data and override the metadata before you load it into a dataset. The dataflow runs on a daily schedule to continually refresh the data.

Dataset Builder

Use the dataset builder to create a single dataset based on data from one or more related Salesforce objects. With the dataset builder you simply point and click to identify and select related Salesforce objects.

Wave Connector for Excel Data

The Salesforce Wave Connector makes it easy to import data from Microsoft Excel 2013 to Wave Analytics.

Upload External Data from the User Interface

You can use the upload user interface to create a single dataset based on external data. You can upload an external data file in a

.csv, .gz, or .zip format. To refresh the data, you can overwrite the data in the dataset by uploading a new external data file.

External Data API

You can use the External Data API to create a single dataset based on external data in the .csv format. You can also use the API to edit the dataset by uploading a new .csv file. When you edit the dataset, you can choose to overwrite all records, append records, update records, or delete records.

Numeric-Value Handling in Datasets

Wave Analytics internally stores numeric values in datasets as long values. For example, it stores the number “3,200.99” with a scale of

“2” as “320099”. The user interface converts the stored value back to decimal notation to display the number as “3200.99.”

The maximum numeric value that can be stored in a dataset is 36,028,797,018,963,967 and the minimum numeric value is

-36,028,797,018,963,968.

Warning: If a numeric value is not within this range, you might receive unexpected results. For example, if you try to load the value 3.7E-16 with a scale of 16 into a dataset, Wave Analytics tries to store the value as 37000000000000000. However, because this value exceeds the maximum, Wave Analytics fails to load the entire record. In addition, if you perform a query that aggregates measures—like sum or group by—and the resulting value exceeds the maximum, the value overflows and Wave Analytics returns an incorrect result.

Date Handling in Datasets

CreateDate

field, Wave Analytics generates date fields such as

CreateDate_Day

and

CreateDate_Week

. If your fiscal year differs from the calendar year, you can enable Wave Analytics to generate fiscal date fields as well.

Wave Analytics generates the following date fields.

Field Name

<date field name>_Second

<date field name>_Minute

Field Type

Text

Description

Number of seconds. If the date contains no seconds, value is '0.'

Number of minutes. If the date contains no minutes, value is '0.'

Analytics Data Integration

Field Name

<date field name>_Hour

<date field name>_Day

<date field name>_Week

<date field name>_Month

<date field name>_Quarter

<date field name>_Year

<date field name>_Week_Fiscal

<date field name>_Month_Fiscal

<date field name>_Quarter_Fiscal

<date field name>_Year_Fiscal

<date field name>_sec_epoch

<date field name>_day_epoch

Field Type

Text

Numeric

Description

Number of hours. If the date contains no hours, value is '0.'

Day of the month.

Week number in calendar year.

Month number in calendar year.

Quarter number in calendar year.

Calendar year.

Week number in fiscal year.

Month number in fiscal year.

Quarter number in fiscal year.

Fiscal year.

Number of seconds that have elapsed since

January 1, 1970 (midnight UTC).

Number of days that have elapsed since

January 1, 1970 (midnight UTC).

You can set metadata attributes to control how dates are loaded into datasets and to enable Wave Analytics to generate fiscal date fields. You set the metadata attributes in the

sfdcDigest transformation parameters

for Salesforce data or in the metadata file for external data.

Important:

Before loading dates from an external data file, ensure that you review the date format requirements here . Also, ensure that the column names in the external data file do not conflict with the generated date field names. For example, if you load a CSV with column

Create_Date

, Wave Analytics generates the

Create_Date_Year

field in the dataset. If the CSV also had a field named

Create_Date_Year

, Wave Analytics would throw an error because the names conflict.

Fiscal Periods in Wave Analytics

If the calendar and fiscal year differ, you can enable Wave Analytics to generate the fiscal date fields in the dataset in addition to calendar date fields. To enable Wave Analytics to generate fiscal date fields, set the fiscalMonthOffset

attribute to a value other than '0'.

You set this attribute for each date column for which you want to generate fiscal date fields. If you set the offset to '0' or you do not specify a value, Wave Analytics does not generate any fiscal date fields.

Additionally, to configure the fiscal periods, set the following metadata attributes for each date column:

fiscalMonthOffset

In addition to enabling the generation of fiscal date fields, this attribute also determines the first month of the fiscal year. You specify the difference between the first month of the fiscal year and first month of the calendar year (January) in fiscalMonthOffset

For example, if your fiscal year begins in April, set fiscalMonthOffset

to '3'.

isYearEndFiscalYear

Because the fiscal year can start in one calendar year and end in another, you must specify which year to use for the fiscal year. The isYearEndFiscalYear

attribute indicates whether the fiscal year is the year in which the fiscal year ends or begins.

Analytics Data Integration

To see how this works, let’s look at a couple of examples. If isYearEndFiscalYear

= true (or you do not specify this attribute), then the fiscal year is the year in which the fiscal year ends. As shown in the following diagram, any dates between 4/1/2015 and

3/31/2016 are part of the fiscal year 2016 because the fiscal year ends in 2016.

If isYearEndFiscalYear

= false, then the fiscal year is the year in which the fiscal year begins. As shown in the following diagram, any dates between 4/1/2015 and 3/31/2016 are part of the fiscal year 2015 because the fiscal year begins in 2015.

Week Numbering in Wave Analytics

For each date loaded into a dataset, Wave Analytics generates the corresponding week number for the calendar year and, if applicable, fiscal year. Similar to the SOQL function

WEEK_IN_YEAR

, week 1 in Wave Analytics is January 1 - January 7. (This is different from the

UTC week()

calculation.)

If needed, you can configure the week to start on a particular day of the week by setting the firstDayOfWeek

attribute. For example, if January 1 is a Saturday and you configure the week to start on a Monday, then week 1 is January 1 - 2. Week 2 starts on Monday,

January 3. Week 3 starts January 10, the following Monday. Notice that week 1 can be a short week to ensure that the subsequent weeks start on the specified day of the week.

Dataflow JSON

Wave Analytics provides a default dataflow that contains some sample transformation logic. This dataflow is just a sample that you must configure before running it.

Analytics Data Integration

To configure the dataflow, you add transformations to the dataflow definition file. A dataflow definition file is a JSON file that contains transformations that represent the dataflow logic. You can add transformations to determine what data to extract, how to transform datasets, and which datasets to register to make available for queries.

After you configure the dataflow, you upload the new dataflow definition file to Wave Analytics.

By default, the dataflow doesn’t run automatically. To start running the dataflow on the schedule, you must manually start the dataflow first. After the first job runs, the dataflow job runs on the daily schedule. The dataflow runs on a daily schedule to capture the latest changes to Salesforce data and changes in the dataflow logic.

You can also stop, reschedule, and monitor dataflow jobs.

SEE ALSO:

Create Datasets with a Dataflow

Dataset Builder

After you select the data to include in the dataset, the dataset builder generates and appends the associated JSON to the dataflow definition file. The dataset is created the next time the dataflow runs. The data in the dataset refreshes each time the dataflow runs.

Wave Connector for Excel Data

The Salesforce Wave Connector makes it easy to import data from Microsoft Excel 2013 to Wave Analytics.

The Wave Connector is available as an app for Excel 2013 on the desktop and Excel Online in Office 365. The Connector is available as an app from the Microsoft Apps for Office store or your organization’s private app catalog. After you install the Connector just point and click to import data from Excel to Salesforce.

IN THIS SECTION:

Install the Wave Connector Excel App

The Wave Connector app gives users a fast, easy way to import data from Excel 2013 into Salesforce Wave Analytics..

Upload External Data from the User Interface

You can use the upload user interface to create a single dataset based on external data. You can upload an external data file in a .csv,

.gz, or .zip format. To refresh the data, you can overwrite the data in the dataset by uploading a new external data file.

When Wave Analytics loads any data into a dataset, it also adds metadata about each column of data. For example, metadata can include the field type, precision, scale, and default value.

For external data, Wave Analytics infers metadata about each column of data in the external data file unless you specify different metadata attributes in a metadata file. A metadata file is a JSON file that describes the structure of an external data file. For example, you can use a metadata file to explicitly set the field type and default value for a specific column of external data. If no metadata file is provided when you upload external data, Wave Analytics treats every column as a dimension and sets the field type to 'Text.' This impacts the type of queries that can be placed on the dataset because you can’t perform mathematical calculations on dataset columns with a Text field type. You can only perform mathematical calculations on dataset columns with a Numeric field type.

Analytics Data Integration

After you create a dataset based on an external data file, you can edit the dataset to apply a new metadata file. This enables you to change the metadata attributes of each column.

Note: Wave temporarily stores the uploaded CSV and metadata files for processing only. After a dataset is created, Wave purges the files.

SEE ALSO:

Create a Dataset with External Data

External Data API

For more information about the External Data API, see the

Wave Analytics External Data API Developer Guide

Create Datasets with a Dataflow

You can use a dataflow to create one or more datasets based on data from Salesforce objects or existing datasets.

IN THIS SECTION:

Design the Dataflow

Before you start creating the dataflow definition file in the .json format, think about the dataflow design. Consider what data to make available for queries, where to extract the data from, and whether you need to transform the extracted data to get the data you want.

Start and Stop a Dataflow

Configure the dataflow based on your dataflow design. You can configure the dataflow to extract data, transform datasets based on your business requirements, and register datasets that you want to make available for queries. To configure the dataflow, add transformations to the dataflow definition file.

You can manually start a dataflow job to load the data into datasets immediately. You can also stop the job while it’s running. You can run a maximum of 24 dataflow jobs during a rolling 24-hour period.

Monitor a Dataflow Job

Use the data monitor to monitor dataflow jobs to ensure that they complete successfully or to troubleshoot them if they fail.

Reschedule and Unschedule a Dataflow

After you run a dataflow job for the first time, it runs on a daily schedule, by default. You can change the time of the daily schedule or schedule it to run on an hourly interval. You might change the schedule to ensure that the data is available by a particular time or to run the job during non-business hours. You can also unschedule a dataflow.

Design the Dataflow

To illustrate some key design decisions, let’s consider an example. In this example, the goal is to create a dataset called “Won Opportunities.”

The dataset will contain opportunity details, including the account name for each opportunity.

Analytics

To create this dataset, you design the following dataflow:

Data Integration

The dataflow extracts opportunity data from the Opportunity object and extracts the account name from the Account object. For each extracted object, the dataflow creates a new dataset.

The dataflow then transforms the datasets created from the extracted data. First, the dataflow joins the opportunity and account data into a new dataset. Next, the dataflow filters the records based on the opportunity stage so that the dataset contains only won opportunities.

Each time the dataflow transforms a dataset, it creates a new dataset.

Finally, because you want users to be able to query won opportunities only, you configure the dataflow to register the final dataset only.

However, if you wanted, you could register any dataset created by the dataflow and register as many datasets as you like.

Carefully choose which datasets to register because:

•

The total number of rows in all registered datasets cannot exceed 250 million per platform license.

•

Users that have access to registered datasets can query their data. Although, you can apply row-level security on a dataset to restrict access to records.

Configure the Dataflow

A dataflow definition file is a JSON file that contains transformations that represent the dataflow logic. The dataflow definition file must be saved with UTF-8 encoding.

Before you can configure a dataflow to process external data, you must upload the external data to Wave Analytics

In Wave Analytics, click the gear icon ( ) and then click Data Monitor to open the data monitor.

The Jobs view of the data monitor appears by default.

2. Select Dataflow View.

3. To download the dataflow definition file, click Download in the actions list (1).

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

For example, based on the design in the previous step , you can add the following transformations:

USER PERMISSIONS

To edit the dataflow definition file:

•

“Edit Wave Analytics

Dataflows”

Analytics Data Integration

4. Make a backup copy of the dataflow definition file before you modify it.

Wave Analytics doesn’t retain previous versions of the file. If you make a mistake, you can upload the previous version to roll back your changes.

5. Add transformations to the dataflow definition file.

{

" Extract_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "CloseDate" },

{ "name": "AccountId" },

{ "name": "OwnerId" }

]

}

" Extract_AccountDetails": {

"action": "sfdcDigest",

"parameters": {

"object": "Account",

"fields": [

{ "name": "Id" },

{ "name": "Name" }

]

}

" Transform_Augment_OpportunitiesWithAccountDetails": {

"action": "augment",

"parameters": {

"left": "Extract_Opportunities",

"left_key": [ "AccountId" ],

"relationship": "OpptyAcct",

"right": "Extract_AccountDetails",

"right_key": [ "Id" ],

"right_select": [

Analytics Data Integration

}

"Name"

]

}

" Transform_Filter_Opportunities": {

"action": "filter",

"parameters": {

"filter": "StageName:EQ:Closed Won",

"source": "Transform_Augment_OpportunitiesWithAccountDetails"

}

" Register_Dataset_WonOpportunities": {

"action": "sfdcRegister",

"parameters": {

"alias": "WonOpportunities",

"name": "WonOpportunities",

"source": "Transform_Filter_Opportunities"

}

Note: The JSON keys and values are case-sensitive. Each bolded key in the previous example JSON contains one action, which identifies the transformation type. The order in which you add the transformations to the dataflow definition file doesn’t matter.

Wave Analytics determines the order in which to process the transformations by traversing the dataflow to determine the dependencies among them.

6. Before you save the dataflow definition file, use a JSON validation tool to verify that the JSON is valid.

An error occurs if you try to upload the dataflow definition file with invalid JSON. You can find JSON validation tool on the internet.

7. Save the dataflow definition file with UTF-8 encoding, and then close the file.

8. In the Dataflow view of the data monitor, click Upload from the action list (1) to upload the updated dataflow definition file.

Note:

Uploading the dataflow definition file does not affect any running dataflow jobs and does not automatically start the dataflow job.

You can now start the dataflow on demand or wait for it to run on the schedule. Users cannot query the registered datasets until the dataflow runs.

If you have enabled replication, you can create multiple dataflows, in addition to the default dataflow, and configure them in the same

way. Look out for the Create button in the data monitor after you enable replication. See Optimize Your Dataflows with Replication .

Analytics Data Integration

Note: Your org can have up to 10 dataflows. This includes the default dataflow, and Wave app dataflows such as Sales Wave and

Service Wave. Keep in mind that each dataflow run counts towards your limit of 24 dataflow runs per day.

Start and Stop a Dataflow

Note: By default, the dataflow doesn’t run automatically. To start running the dataflow on the schedule, you must manually start the dataflow first. After the first job runs, the dataflow job runs on the daily schedule.

In Wave Analytics, click the gear icon ( monitor.

) and then click Data Monitor to open the data

The Jobs view of the data monitor appears by default.

2. Select Dataflow View.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

USER PERMISSIONS

To start a dataflow job:

•

“Edit Wave Analytics

Dataflows”

Analytics Data Integration

3. Click Start in the actions list (1) to start the dataflow job.

The dataflow job is added to the job queue. The Start button is greyed out while the dataflow job runs.

4. After the job completes, Wave Analytics sends an email notification to the user who last modified the dataflow definition file.

The email notification indicates whether the job completed successfully. It also shows job details like start time, end time, duration, and number of processed rows. If the job failed, the notification shows the reason for the failure.

To stop a dataflow job that is currently running, click next to the job status.

If you click Start to restart a stopped dataflow, the job starts over—the dataflow job does not resume from the point at which it was stopped.

You can monitor the dataflow job in the data monitor to determine when dataflow completes. After the dataflow completes successfully, refresh the Home page to view the registered datasets.

Monitor a Dataflow Job

Use the data monitor to monitor dataflow jobs to ensure that they complete successfully or to troubleshoot them if they fail.

The Dataflow view of the data monitor shows the status, start time, and duration of the last 10 dataflow jobs and retains the last 7 days of history. To help you troubleshoot a failed job, you can view error messages about the job, view the run-time details about every transformation that is processed, and download error logs where available.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

USER PERMISSIONS

To access the data monitor:

•

“Edit Wave Analytics

Dataflows”, “Upload

External Data to Wave

Analytics”, or “Manage

Wave Analytics”

To download an error log:

•

“Edit Wave Analytics

Dataflows” and “View All

Data”

100

Analytics Data Integration

In Wave Analytics, click the gear button (

The Jobs view shows by default.

2. Select Dataflow View (1).

), and then click Data Monitor to open the data monitor.

Click to see the latest status of a job.

Each job can have one of the following statuses:.

Status

Running

Failed

Successful

Warning

Description

The job is running.

The job failed.

The job completed successfully.

The job completed successfully, but some rows failed.

4. If the dataflow job fails, expand the job node (2) and review the run-time details for every transformation that was processed.

5. If an error log is available for a node, click the download log button (3) to download a CSV file containing the failed rows.

Note: Error logs display the data from rows that have failed to load. To maintain data security and prevent unauthorized access to this data, only users with the "View All Data" permission can download error logs.

6. If there’s a problem with the dataflow logic, edit the dataflow definition file, upload it, and then run the dataflow again.

101

Analytics Data Integration

Reschedule and Unschedule a Dataflow

In Wave Analytics, click the gear icon ( monitor.

The Jobs view shows by default.

) and then click Data Monitor to open the data

2. Select Dataflow View.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

computeExpression Transformation

USER PERMISSIONS

To schedule or unschedule a dataflow job:

•

“Edit Wave Analytics

Dataflows”

3. Click Schedule in the actions list (1) to reschedule the dataflow job.

The default daily schedule and start time appears.

4. To continue running it once per day, select the time that you’d like to start the dataflow job.

5. To run it on an hourly basis, select Hour(s) and then specify the interval details. Minute offset is the number of minutes past the hour to run the dataflow.

102

Analytics Data Integration

Hourly scheduled dataflows run relative to a fixed base time (00:00 UTC). For example, if you schedule a dataflow to run every 6 hours with a 15-minute offset, the dataflow is scheduled to run at 00:15 UTC, 06:15 UTC, 12:15 UTC, and 18:15 UTC, each day. The first dataflow run occurs at the next scheduled time. So, if you change the schedule at 04:00 UTC, the dataflow runs at the next scheduled time (06:15 UTC), and then every 6 hours thereafter. To keep things simple, the scheduler specifies the time of the first dataflow run (in your time zone), after you save the schedule.

6. Click Save.

The Dataflow Schedule pop-up box specifies the next time, in your local time zone, that the dataflow runs.

7. Click Done.

8. To stop a dataflow from running on a schedule, select Unschedule in the action list (1).

Transformations for Wave Dataflows

For example, you can use transformations to join data from two related datasets and then register the resulting dataset to make it available for queries.

IN THIS SECTION:

append Transformation

The append transformation combines records from multiple datasets into a single dataset.

augment Transformation

The augment transformation augments an input dataset by combining columns from another related dataset. The resulting, augmented dataset enables queries across both related input datasets. For example, you can augment the Account dataset with the User dataset to enable a query to return account records and the full names of the account owners.

The computeExpression transformation enables you to add derived fields to a dataset. The values for derived fields aren’t extracted from the input data source. Instead, Wave generates the values using a SAQL expression, which can be based on one or more fields from the input data or other derived fields. For example, you can use an expression to assign a value to a field, concatenate text fields, or perform mathematical calculations on numeric fields.

103

Analytics Data Integration

computeRelative Transformation

You can use the computeRelative transformation to analyze trends in your data by adding derived fields to a dataset based on values in other rows. For example, to analyze sales pipeline trends, create derived fields that calculate the number of days an opportunity remains in each stage. You can also calculate the changes to the opportunity amount throughout the stages of the opportunity.

delta Transformation

The delta transformation calculates changes in the value of a measure column in a dataset over a period of time. The delta transformation generates an output column in the dataset to store the delta for each record. Create deltas to make it easier for business analysts to include them in queries.

dim2mea Transformation

The dim2mea transformation creates a new measure based on a dimension. The transformation adds the new measure column to the dataset. The transformation also preserves the dimension to ensure that existing lenses and dashboards don’t break if they use the dimension.

edgemart Transformation

The edgemart transformation gives the dataflow access to an existing, registered dataset, which can contain Salesforce data, external data, or a combination of the two. Use this transformation to reference a dataset so that its data can be used in subsequent transformations in the dataflow. You can use this transformation and the augment transformation together to join an existing dataset with a new dataset.

filter Transformation

The filter transformation removes records from an existing dataset. You define a filter condition that specifies which records to retain in the dataset.

flatten Transformation

The flatten transformation flattens hierarchical data. For example, you can flatten the Salesforce role hierarchy to implement row-level security on a dataset based on the role hierarchy.

sfdcRegister Transformation

The sfdcDigest transformation generates a dataset based on data that it extracts from a Salesforce object. You specify the Salesforce object and fields from which to extract data. You might choose to exclude particular fields that contain sensitive information or that aren’t relevant for analysis.

The sfdcRegister transformation registers a dataset to make it available for queries. Users cannot view or run queries against unregistered datasets.

sliceDataset Transformation

The sliceDataset transformation removes fields from a dataset in your dataflow, leaving you with a subset of fields for use in a new dataset or in other transformations. This allows you to create multiple datasets, each with different sets of fields from a single dataset.

update Transformation

The update transformation updates the specified field values in an existing dataset based on data from another dataset, which we’ll call the lookup dataset. The transformation looks up the new values from corresponding fields in the lookup dataset. The transformation stores the results in a new dataset.

Overriding Metadata Generated by a Transformation

Optionally, you can override the metadata that is generated by a transformation. You can override object and field attributes. For example, you can change a field name that is extracted from a Salesforce object so that it appears differently in the dataset. To override the metadata, add the overrides to the Schema section of the transformation in the dataflow definition file.

SEE ALSO:

Create Datasets with a Dataflow

104

Analytics Data Integration

append Transformation

The append transformation combines records from multiple datasets into a single dataset.

Consider the following rules when using this transformation.

•

This transformation does not remove duplicate records.

•

All input datasets must have the same structure—the corresponding columns must be in the same order and have the same name and field type.

Example:

Let’s look at an example. Each month, you create a dataset that contains the month’s sales targets. Now, you want a holistic view of sales targets for all months. To do that, you create the following dataflow definition file to merge the existing datasets into a single dataset.

{

"Extract_SalesTargets_2014Jan": {

"action": "edgemart",

"parameters": { "alias": "SalesTargets_2014Jan" }

"Extract_SalesTargets_2014Feb": {

"action": "edgemart",

"parameters": { "alias": "SalesTargets_2014Feb" }

"Extract_SalesTargets_2014Mar": {

"action": "edgemart",

"parameters": { "alias": "SalesTargets_2014Mar" }

"Append_SalesTargets_2014Quarter1": {

"action": "append",

"parameters": {

"sources": [

"Extract_SalesTargets_2014Jan",

"Extract_SalesTargets_2014Feb",

"Extract_SalesTargets_2014Mar"

]

}

"Register_AllSalesTargets": {

"action": "sfdcRegister",

"parameters": {

"alias": "AllSalesTargets",

"name": "AllSalesTargets",

"source": "Append_SalesTargets_2014Quarter1"

}

After you create the single dataset, you can use date filters to analyze the sales targets by month, quarter, or year.

105

Analytics Data Integration

IN THIS SECTION:

append Parameters

When you define an append transformation, you set the action attribute to append

and specify the parameters.

append Parameters

When you define an append transformation, you set the action attribute to append

and specify the parameters.

The following table describes the input parameters.

Parameter

sources

Required?

Yes

Value

Nodes in the dataflow definition file that identify the datasets that you want to merge.

SEE ALSO:

append Transformation

augment Transformation

When you create the transformation, you identify each input dataset as the left or right dataset and specify the relationship between them. Wave Analytics combines all the columns of the left dataset with only the specified columns from the right dataset. (Keep in mind that each dataset can’t have more than 5,000 columns.) Wave Analytics adds the relationship to column names from the right dataset, which is useful when the left and right datasets have columns with the same names.

For each record in the left dataset, the augment transformation performs a lookup to find a matching record in the right dataset. To match related records, the augment transformation uses a match condition. You specify the match condition based on a key from each dataset. For example:

"left_key": [ "Id" ],"right_key": [ "Prod_ID" ]

A key can be a single-column key or a composite key. For a match condition based on a composite key, the keys for both datasets must have the same number of columns, specified in the same order.

Tip: To augment more than two datasets, augment two datasets at a time. For example, to augment three datasets, augment the first two datasets, and then augment the resulting dataset with the third dataset.

Example: Let’s look at an example of the augment transformation. In this example, you want to extract data from the Opportunity and Accounts objects, and then match the data based on the account ID field.

106

Analytics

You create the following dataflow definition file.

{

"Extract_Opportunity": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "CloseDate" },

{ "name": "AccountId" },

{ "name": "OwnerId" }

]

}

"Extract_AccountDetails": {

"action": "sfdcDigest",

"parameters": {

"object": "Account",

"fields": [

{ "name": "Id" },

{ "name": "Name" }

]

}

"Augment_OpportunitiesWithAccountDetails": {

"action": "augment",

"parameters": {

"operation": "LookupSingleValue",

"left": "Extract_Opportunity",

"left_key": [

"AccountId"

"relationship": "OpptyAcct",

"right": "Extract_AccountDetails",

"right_key": [

"Id"

"right_select": [ "Name" ]

}

107

Data Integration

Analytics Data Integration

"Register_OpportunityDetails": {

"action": "sfdcRegister",

"parameters": {

"alias": "Opportunity_Account",

"name": "Opportunity_Account",

"source": "Augment_OpportunitiesWithAccountDetails" }

}

After you run the dataflow, Wave Analytics creates and registers the Opportunity_Account dataset. It also adds the relationship as a prefix to all columns from the right dataset.

IN THIS SECTION:

Special Cases for Matching Records with the augment Transformation

For each record in the left dataset, the augment transformation performs a lookup to find a matching record in the right dataset.

However, it’s critical that you understand how the augment transformation handles special cases when matching records.

augment Parameters

When you define an augment transformation, you set the action attribute to augment

and specify the parameters.

Special Cases for Matching Records with the augment Transformation

For each record in the left dataset, the augment transformation performs a lookup to find a matching record in the right dataset. However, it’s critical that you understand how the augment transformation handles special cases when matching records.

Let’s look at some examples that illustrate some special cases.

Handling Null Keys

When a record in the left dataset contains a null key, Wave Analytics doesn’t perform a lookup to match the record. Instead, Wave

Analytics appends the right columns and inserts null for dimensions (including dates) and '0' for measures.

Let’s look at an example. You apply the augment transformation on the following datasets, set the relationship to "Price,", and match the records based on the

and

ProdId

fields.

108

Analytics Data Integration

Wave Analytics doesn’t match the last record because the product ID is null. Instead, Wave Analytics inserts a null for the

Price.Pricebook

dimension and '0' for the

Price.UnitPrice

measure. Here’s the resulting dataset after the augment.

Handling Empty Keys

Wave Analytics matches empty-value left keys and empty-value right keys.

Let’s look at an example. You apply the augment transformation on the following datasets, set the relationship to "Price,", and match the records based on the

and

ProdId

fields.

Wave Analytics matches the last record in the Product dataset with the third record in the Price dataset because they both have empty values (""). Here’s the resulting dataset after the augment.

Handling Non-Unique Keys

Although it’s recommended, the left key doesn’t have to be unique. If multiple records have the same left key, Wave Analytics creates the same values for the appended columns.

Let’s look at an example. You apply the augment transformation on the following datasets, set the relationship to "Price,", and match the records based on the

and

ProdId

fields.

109

Analytics Data Integration

Wave Analytics matches the records in the Product dataset with records in the Price dataset. Here’s the resulting dataset after the augment.

Handling No Match

If left key doesn't have a match in the right dataset, Wave Analytics appends the right columns and inserts null for dimensions (including dates) and '0' for measures.

Let’s look at an example. You apply the augment transformation on the following datasets, set the relationship to "Price,", and match the records based on the

and

ProdId

fields.

Because no keys match, Wave Analytics doesn’t match any records in the Product dataset with records in the Price dataset. Here’s the resulting dataset after the augment.

Handling Multiple Matches

If the left dataset has a one-to-many relationship with the right dataset, Wave Analytics might find multiple matches for a left record.

What Wave Analytics does with multiple matches depends on the specified augment operation. You can specify one of the following operations to handle the multiple-match case:

LookupSingleValue

The augment transformation returns results from a single row. Wave Analytics randomly selects one row from the list of matched rows.

Note:

Each time you run the dataflow, Wave Analytics can return different results depending on the returned row.

110

Analytics Data Integration

Let’s look at an example. You apply the augment transformation on the following datasets, set the relationship to "Price," set the operation to

LookupSingleValue

, and match the records based on the

and

ProdId

fields.

Although there are multiple rows for Prod3 in the Price dataset, Wave Analytics randomly chooses one matching row and returns values based on that row. Here’s the resulting dataset after the augment if Wave Analytics chooses the first Prod3 row.

LookupMultiValue

Wave Analytics returns results from all matched rows.

Let’s look at an example. You apply the augment transformation on the following datasets, set the relationship to "Price," set the operation to

LookupMultiValue

, and match the records based on the

and

ProdId

fields.

Because the lookup returns multiple rows for Prod3, the dimension

Price.Pricebook

field in the resulting dataset becomes a multi-value field, showing all dimension values. The measure field

Price.UnitPrice

contains 1500, which is the sum of 800 and 700. Here’s the resulting dataset after the augment.

SEE ALSO:

augment Transformation

augment Parameters

When you define an augment transformation, you set the action attribute to augment

and specify the parameters.

The following table describes the input parameters.

111

Analytics

Parameter

operation

Required?

No left left_key right relationship right_select

Yes

Data Integration

Value

Indicates what the transformation does if it matches multiple rows in the right dataset with a row in the left. Valid values:

•

LookupSingleValue

. Returns values from one of the matched rows.

If you don’t specify the operation parameter, the transformation uses this operation.

•

LookupMultiValue

. Returns values from all matched rows.

For more information about each operation,

see Special Cases for Matching Records with the augment Transformation

Node in the dataflow definition file that identifies the left dataset. This is one of two input sources for this transformation.

Key column in the left dataset used to augment the datasets. If you use a composite key, the left and right keys must have the same number of columns in the same order. For a composite key, use the following syntax:

[ "Key Column1", "Key

Column2", …, "Key ColumnN" ]

Note:

The left or right key can’t be a multi-value field.

Node in the dataflow definition file that identifies the right dataset. This is one of two input sources for this transformation.

Relationship between the left and right datasets. The dataflow adds the relationship to the beginning of the right column names in the output dataset to make the column names unique and descriptive.

An array of column names from the right dataset that you want to include in the output dataset. The dataflow adds the relationship as a prefix to the column name to determine the name of the right column in the output dataset.

112

Analytics Data Integration

Parameter

right_key

Required?

Yes

Value

Key column in the right dataset used to augment the datasets. If you use a composite key, the left and right keys must have the same number of columns in the same order.

SEE ALSO:

augment Transformation

computeExpression Transformation

Note: The computeExpression and computeRelative transformations are similar, but they have a key difference. The computeExpression transformation performs calculations based on other fields within the same row. The computeRelative transformation performs calculations based on the previous and next values of the same field in other rows.

Consider the following guidelines when creating a computeExpression transformation:

•

You can include only the following SAQL operators and functions in the expression:

–

Arithmetic operators

–

Case operator

–

String operator

–

Date functions

•

The values of the derived field must match its specified type. For example, set the type of the derived field to

Text

if the values are strings.

•

Wave calculates the values of derived fields in the order in which they are listed in the JSON. Thus, if you create a derived field based on other derived fields in the same computeExpression transformation, the input derived fields must be listed first. For example,

Derived_A must be listed before Derived_B in the following computeExpression transformation JSON snippet:

"CreateDerivedFields": {

"action": "computeExpression",

"parameters": {

"source": "sourceNode",

"mergeWithSource": false,

"computedFields": [

{

"name": "Derived_A",

"type": "Text",

"label": "Derived Field A",

"saqlExpression": "\"hello \""},

{

"name": "Derived_B",

"type": "Text",

"label": "Derived Field B Dependent on Field A",

113

Analytics Data Integration

]

"saqlExpression": "Derived_A + \"world\""}

}

•

You can choose whether the resulting dataset includes only the derived fields, or includes the input and derived fields.

Example: Let’s look at an example. You want to create a dataset based on Salesforce opportunity data. You create a dataflow that extracts the Id and Amount fields from the Opportunity object. In addition, you also want to add the following derived fields to the dataset: ModifiedId, SalesTax, FinalPrice, and ValueCategory. For the derived fields, you will:

•

Append “SFDC” to each opportunity Id to get a new modified Id.

•

Calculate the sales tax based on an 8% tax rate.

•

Calculate the final price by adding the amount and sales tax.

•

Categorize opportunities into low-, medium-, and high-value buckets based on the calculated final price.

You create the following dataflow definition.

{

"salesData": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{"name": "Amount"},

{"name": "Id"}]}},

"Derived_Fields": {

"action": "computeExpression",

"parameters": {

"source": "salesData",

"mergeWithSource": true,

"computedFields": [

{

"name": "ModifiedId",

"type": "Text",

"saqlExpression": "\"SFDC\" + Id"},

{

"name": "SalesTax",

"type": "Numeric",

"precision": 18,

"defaultValue": "0",

"scale": 5,

"saqlExpression": "Amount * 0.08"},

{

"name": "FinalPrice",

"type": "Numeric",

"precision": 18,

"defaultValue": "0",

"scale": 5,

"saqlExpression": "Amount + SalesTax"},

{

"name": "ValueCategory",

"type": "Text",

"saqlExpression": "

114

Analytics Data Integration

}

]

}

"Register_CategorizedSales": {

"action": "sfdcRegister",

"parameters": {

"alias": "Categorized_Sales",

"name": "Categorized_Sales",

"source": "Derived_Fields" }

} case when FinalPrice < 1000 then \"Low\" when FinalPrice >= 1000 and FinalPrice < 2000 then \"Medium\" else \"High\" end"}

IN THIS SECTION:

computeExpression Parameters

When you define a computeExpression transformation, you set the action attribute to computeExpression

. You also specify the parameters for the input source and the expression used to generate the values.

computeExpression Parameters

When you define a computeExpression transformation, you set the action attribute to computeExpression

. You also specify the parameters for the input source and the expression used to generate the values.

You can specify parameters in the following sections of the computeExpression node: parameters and computedFields.

Parameters

The following table describes the parameters in the parameters

section.

Parameter

source

Required?

Yes mergeWithSource No

Value

Node in the dataflow definition file that identifies the input source for this transformation.

Indicates whether the input fields are included with the derived fields in the resulting dataset. When true, the resulting dataset contains all input fields from the source and the newly generated derived fields. When false, the resulting dataset contains the derived fields only. Default is true.

115

Analytics

Parameter

computedFields

Required?

Yes

Data Integration

Value

Attributes and expression used to generate derived fields in the dataset. See

computedFields

The following table describes the attributes in the computedFields

section. It also describes optional attributes that you can provide to override the field metadata to make the data appear differently in a dataset. For example, Wave can replace null values in a field with a default value.

Parameter

name

Required?

Yes

Value

API name of the generated field.

Note:

The API names must be unique. Otherwise, the dataflow fails to run.

type Yes

Wave Analytics field type associated with the field. Valid types are Text, Numeric, or

Date.

Example:

"type": "Text" label saqlExpression

Yes

The display name of the generated field that appears in the Wave user interface. Can be up to 255 characters. Defaults to input field name if not specified.

SAQL expression used to calculate the value for the derived field. The expression can be based on input fields or other derived fields in the transformation.

Example:

"saqlExpression":"toDate(birth_day,

\"yyyy-M-d\")" format precision

Yes (for Date fields only)

Yes (for Numeric fields only)

Format of the derived date field. For information about formats, see the External

Data Format Reference .

The maximum number of digits in a numeric value, or the length of a text value. For numeric values: Includes all numbers to the left and to the right of the decimal point

116

Analytics

Parameter

scale defaultValue

Required?

Yes (for Numeric fields only)

Data Integration

Value

(but excludes the decimal point character).

Value must be between 1 and 18, inclusive.

For text values: Value defaults to 255 characters, and must be between 1 and

32,000 characters, inclusive.

Example:

"precision": 10

The number of digits to the right of the decimal point in a numeric value. Must be less than the precision value. Value must be between 0 and 17 characters, inclusive.

Example:

"scale": 2

For text and numeric fields that can be null.

Default value that replaces a null value for the specified field.

SEE ALSO:

computeExpression Transformation

computeRelative Transformation

Note: The computeExpression and computeRelative transformations are similar, but the computeExpression transformation performs calculations based on fields within the same row. The computeRelative transformation performs calculations based on the same field in other rows—particularly the current, first, previous, or next rows.

When you define a computeRelative transformation, you specify a source transformation as the input, partition the records, and sort the records within each partition. For example, you can use sfdcDigest to extract opportunity history records, and then use computeRelative to calculate changes in each opportunity over time. You can partition opportunity records by opportunity ID, and then chronologically sort records within each partition to correctly identify the previous and next values.

Example: Let’s look at an example. To perform trending analysis on the sales pipeline, create a dataflow that contains the following transformations.

sfdcDigest transformation

Extracts the following data from the OpportunityHistory object.

117

Analytics Data Integration

computeRelative transformation

Performs the following tasks:

•

Partitions the extracted records by opportunity ID.

•

Within each partition, sorts the extracted records by CreatedDate in ascending order. Sorting by CreatedDate ensures that the changes that occur for each opportunity are listed in chronological order.

•

Adds the following derived fields to the final dataset.

OpportunityCreatedDate

Determines the date that the opportunity was first created. You can use this date with the actual close date to determine the number of days required to close the sale. The goal is to shorten the sales cycle to recognize revenue.

ValidToDate

Determines the date to which the state of the opportunity is valid. You can use this field to determine the time period for which the opportunity details are valid in each record.

AmountPrev

Determines the previous amount of the opportunity. You can use this field to determine if the values of opportunities are increasing or decreasing, which can affect whether you hit your sales targets.

CloseDatePrev

Determines the previous expected close date for the opportunity. You can use this field to analyze how the expected close date changes over the sales cycle of the opportunity. If the expected close date keeps getting pushed out, identify the issues that are causing the longer sales cycle.

sfdcRegister transformation

Registers the final dataset that contains the extracted fields from the sfdcDigest transformation and the derived fields from computeRelative transformation.

You create the following dataflow definition.

{

"extractOppHistory": {

"action": "sfdcDigest",

"parameters": {

"object": "OpportunityHistory",

"fields": [

{"name": "OpportunityId"},

{"name": "CreatedDate"},

{"name": "StageName"},

{"name": "Amount"},

{"name": "CloseDate"}

]

118

Analytics

}

"computeTrending": {

"action": "computeRelative",

"parameters": {

"partitionBy": ["OpportunityId"],

"orderBy": [

{

"name":"CreatedDate",

"direction":"asc"

}

"computedFields": [

{

"name": "OpportunityCreatedDate",

"expression": {

"sourceField": "CreatedDate",

"offset": "first()",

"default": "current()"

}

{

"name": "ValidToDate",

"expression": {

"sourceField": "CreatedDate",

"offset": "next()",

"default": "3000-01-01T00:00:00.000Z"

}

{

"name": "AmountPrev",

"expression": {

"sourceField": "Amount",

"offset": "previous()",

"default": "0"

}

{

"name": "CloseDatePrev",

"expression": {

"sourceField": "CloseDate",

"offset": "previous()",

"default": "01/01/1970"

}

"source": "extractOppHistory"

}

"Register_OppportunityHistory_Dataset": {

"action": "sfdcRegister",

"parameters": {

"alias": "SalesPipelineTrending",

"name": "Sales Pipeline Trending",

119

Data Integration

Analytics

"source": "computeTrending"

}

The dataflow runs and creates the following dataset.

Data Integration

Notice that Wave partitions the records by opportunity ID and then sorts the records in ascending order based on the CreatedDate field within each partition. Wave can now use the previous and next rows within each partition to determine changes in field values in the dataset.

IN THIS SECTION:

computeRelative Parameters

When you define a computeRelative transformation, you set the action attribute to computeRelative

. You also specify the parameters for the input source, partition-by field, sort field, and derived field definitions.

computeRelative Parameters

When you define a computeRelative transformation, you set the action attribute to computeRelative

. You also specify the parameters for the input source, partition-by field, sort field, and derived field definitions.

You can specify parameters in the following sections of the computeRelative node.

Parameters

The following table describes the parameters in the parameters

section.

Parameter

source partitionBy orderBy

Required?

Yes

Value

Node in the dataflow definition file that identifies the input source for this transformation.

API name of the field used to partition the records in the dataset. Specify one partition-by field only.

Field used to sort the records within each partition and the sort order: ascending

( asc

) or descending ( desc

). Specify one sort field only.

120

Analytics

Parameter

Data Integration computedFields

Required?

Yes

Value

Example:

"orderBy": [

{

"name":"CreatedDate",

"direction":"asc"

}

]

A list of definitions for derived fields.

Example:

"computedFields": [

{

"name": "OpportunityCreatedDate",

"expression": {

"sourceField": "CreatedDate",

"offset": "first()",

"default": "current()"

}

{

"name": "AmountPrev",

"expression": {

"sourceField": "Amount",

"offset": "previous()",

"default": "0"

}

]

See computedFields

computedFields

The following table describes the attributes in the computedFields

section.

Parameter

name

Required?

Yes

Value

API name of the derived field to add to the dataset. The name must be unique in the dataset.

Note: If the name is not unique, the dataflow fails to run.

expression Yes

Expression attributes used to calculate the value for the derived field. The expression can be based on input fields or other derived fields in the transformation.

121

Analytics

Parameter

Data Integration

Required?

Value

Example:

"expression": {

"sourceField": "CloseDate",

"offset": "previous()",

"default": "01/01/1970"

See expression

expression

The following table describes the attributes in the expression

section.

Parameter

sourceField offset

Required?

Yes

Value

API name of the input field from the source node that’s used in the expression.

The function used in the expression. You can use the following functions:

current()

Gets the value from the current record.

Example:

"offset": "current()"

first()

Gets the value from the first record in the partition, like the first CreateDate for an opportunity.

Example:

"offset": "first()"

next()

Gets the value from the next record.

Example:

"offset": "next()"

previous()

Gets the value from the previous record.

Example:

"offset": "previous()"

Note:

Derived fields are computed in the order that they’re defined. The calculation of a derived field can be based on the value from another derived field as long as it has already been defined. For example, next() can’t access the value of a derived field in the next row.

122

Analytics

Parameter

default

Data Integration

Required?

Value

Tip: To get the correct results when using the previous()

and next()

functions, the computeRelative transformation requires you to sort the records.

Yes (for numeric fields only)

The default value if one can’t be calculated. For example, you can specify a default value when no previous or next value exists. You can insert a constant value or current()

as a default value.

Examples:

"default": "3000-01-01T00:00:00.000Z"

"default": "current()"

delta Transformation

The delta transformation calculates each delta value by comparing the value in each record with the value in the previous record. Because records might not be sorted, the delta transformation orders the records before computing the delta values. To do this, the transformation sorts the data by the specified dimension, and then by the specified epoch date column.

Note:

When Wave Analytics processes dates, it creates the following epoch date columns for each date processed:

Epoch Time Column

<date_column_name>_sec_epoch

<date_column_name>_day_epoch

Description

For example, if the date column is CloseDate, the generated epoch second column is CloseDate_sec_epoch. This column provides the number of seconds that have elapsed since January

1, 1970 (midnight UTC/GMT).

For example, if the date column is CloseDate, the generated epoch day column is CloseDate_day_epoch. This column provides the number of days that have elapsed since January

1, 1970 (midnight UTC/GMT).

Example:

Let’s look at an example. You want to create an OppHistoryDelta dataset that contains opportunity history from the

OpportunityHistory object and also calculates the deltas for opportunity amounts.

The OpportunityHistory object contains the following data.

OpportunityId

CloseDate

1/1/2014

StageName

New

Amount

100

123

Analytics

OpportunityId

CloseDate

2/1/2014

3/1/2014

StageName

ClosedWon

Amount

200

100

You create the following dataflow definition.

{

"Extract_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "OpportunityHistory",

"fields": [

{ "name": "OpportunityId" },

{ "name": "CloseDate" },

{ "name": "StageName" },

{ "name": "Amount" }

]

}

"Calculate_Delta": {

"action": "delta",

}

"Register_Dataset": {

"action": "sfdcRegister",

"parameters": {

"alias": "OppHistoryDelta",

"name": "OppHistoryDelta",

"source": "Calculate_Delta"

}

"parameters": {

"dimension": "OpportunityId",

"epoch": "CloseDate_day_epoch",

"inputMeasure": "Amount",

"outputMeasure": "DeltaAmount",

"source": "Extract_Opportunities"

}

To calculate the delta values for each opportunity amount, the delta transformation sorts the records by the dimension

(OpportunityId) first, and then by time (CloseDate_day_epoch) as shown here.

OpportunityID

CloseDate

1/1/2014

3/1/2014

1/1/2014

2/1/2014

StageName

New

ClosedWon

New

ClosedWon

Amount

100

200

Data Integration

124

Analytics Data Integration

For the first record of each dimension, the transformation inserts ‘0’ for the delta value.

Note: If an opportunity contains multiple changes on the same day, you must sort the records on a shorter time interval.

In this case, sort on CloseDate_sec_epoch column. Otherwise, records might not be sorted correctly, which means delta values will be incorrect.

IN THIS SECTION:

delta Parameters

When you define a delta transformation, you set the action attribute to delta

and specify the parameters.

delta Parameters

When you define a delta transformation, you set the action attribute to delta

and specify the parameters.

The following table describes the input parameters:

Parameter

dimension epoch inputMeasure outputMeasure

Required?

Yes

Value

Dimension column in the dataset used to sort records when calculating the delta values.

Epoch date column in the dataset used to sort records within each dimension when calculating delta values.

Measure column on which you want to calculate the delta.

Name of the output column that contains the delta value.

125

Analytics Data Integration

Parameter

source

Required?

Yes

Value

Node in the dataflow definition file that contains the dataset to which you want to add the delta column.

SEE ALSO:

delta Transformation

dim2mea Transformation

If the transformation cannot create a measure from a dimension, the transformation populates the measure with the specified default value. If no default value is provided, the transformation inserts ‘0.’

Example: Let’s look at an example. Your Opportunity object contains a custom text field called StageVal__c, which contains the opportunity amount at a particular stage. Because this is a text field, Wave Analytics loads this data as a dimension. However, you’d like to create a measure from this dimension to enable users to perform calculations on stage amount.

You create the following dataflow definition.

{

"Extract_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "CloseDate" },

{ "name": "AccountId" },

{ "name": "StageVal__c" }

]

}

"Create_Measure_From_Dimension": {

"action": "dim2mea",

"parameters": {

"dimension": "StageVal__c",

"measure": "StageValue",

"measureDefault": "0",

"measureType": "long",

"source": "Extract_Opportunities"

}

"Register_The_Dataset": {

"action": "sfdcRegister",

"parameters": {

126

Analytics Data Integration

}

"alias": "OpportunitiesWithConvertedMeasure",

"name": "OpportunitiesWithConvertedMeasure",

"source": "Create_Measure_From_Dimension"

IN THIS SECTION:

dim2mea Parameters

When you define a dim2mea transformation, you set the action attribute to dim2mea

and specify the parameters.

dim2mea Parameters

When you define a dim2mea transformation, you set the action attribute to dim2mea

and specify the parameters.

The following table describes the input parameters:

Parameter

dimension measure measureDefault measureType source

Required?

Yes

Value

Dimension column in the dataset from which you want to create the measure.

Name of the output measure. This column name must be unique within the dataset.

Do not use the same name as the dimension because the transformation preserves the dimension in the dataset.

Default value for the measure if the transformation is unable to create a measure from a dimension.

Type of measure. Valid value: “long”

Node in the dataflow definition file that contains the dataset to which you want to add the measure.

SEE ALSO:

dim2mea Transformation

edgemart Transformation

127

Analytics Data Integration

You would now like to create a dataset that contains opportunity information from the Opportunity object. Then, you would like to join the data from the existing FinalSales dataset with the Opportunity dataset.

You create the following dataflow definition.

{

"Extract_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "Amount" }

]

}

"Extract_Final_Sales_Data": {

"action": "edgemart",

"parameters": { "alias": "FinalSales" }

"Combine_Opportunities_FinalSales": {

"action": "augment",

"parameters": {

"left": "Extract_Opportunities",

"left_key": [ "Id" ],

"relationship": "Opportunity",

"right": "Extract_Final_Sales_Data",

"right_key": [ "OppID" ],

"right_select": [ "SaleAmount" ]

}

"Register_Opportunity_FinalSales_Dataset": {

"action": "sfdcRegister",

"parameters": {

"alias": "OpportunityVersusFinalSales",

"name": "OpporunityVersusFinalSales",

"source": "Combine_Opportunities_FinalSales"

}

128

Analytics Data Integration

IN THIS SECTION:

edgemart Parameters

When you define an edgemart transformation, you set the action attribute to edgemart

and specify the parameters.

edgemart Parameters

When you define an edgemart transformation, you set the action attribute to edgemart

and specify the parameters.

The following table describes the input parameter:

Parameter

alias

Required?

Yes

Value

API name of the dataset from which you want to extract data. To determine the API name of a dataset, edit the dataset and view the system name.

SEE ALSO:

edgemart Transformation

filter Transformation

The filter transformation removes records from an existing dataset. You define a filter condition that specifies which records to retain in the dataset.

Example: Let’s look at an example. You would like to create a dataset that contains only opportunities that were Closed Won.

First, you extract all opportunities from the Opportunity object. Next, you filter the records so that you only include opportunities with a Closed Won stage name.

You create the following dataflow definition.

{

"Extract_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "CloseDate" },

{ "name": "AccountId" },

{ "name": "OwnerId" }

]

}

"Filter_Opportunities": {

"action": "filter",

"parameters": {

"filter": "StageName:EQ:Closed Won",

"source": "Extract_Opportunities"

129

Analytics Data Integration

}

"Register_My_Won_Oppportunities_Dataset": {

"action": "sfdcRegister",

"parameters": {

"alias": "MyWonOpportunities",

"name": "MyWonOpportunities",

"source": "Filter_Opportunities"

}

IN THIS SECTION:

filter Parameters

When you define a filter transformation, you set the action attribute to filter

and specify the parameters.

filter Expression Syntax

You create a filter expression in the filter transformation based on one or more dimensions in a dataset.

filter Parameters

When you define a filter transformation, you set the action attribute to filter

and specify the parameters.

The following table describes the input parameters:

Parameter

filter source

Required?

Yes

Value

Filter expression that specifies which records to include in the new dataset. See

filter

Expression Syntax .

Node in the dataflow definition file that contains the dataset that you want to filter.

SEE ALSO:

filter Transformation

filter Expression Syntax

You create a filter expression in the filter transformation based on one or more dimensions in a dataset.

Note: String comparisons in a filter expression are case-sensitive.

You can use the following types of filter expressions:

Filter Expression Syntax

dim:EQ:value

Description

True if the dimension and value are equal.

Example:

"filter": "StageName:EQ:Closed Won"

130

Analytics

Filter Expression Syntax

dim:R:val0:val1 dim:R:val dim:R::val dim:N:val dim:EQ:val1|val2 dim1:EQ:val1,dim2:EQ:val2

Data Integration

Description

True if the left dimension falls within the specified range between val0 and val1.

Example:

"filter": "EmployeeId:R:100:1000"

True if the dimension is greater than or equal to the value based on binary sort order. For example, this is true when the dimension is 'City' and the value is 'Anderson' because ' 'City' > 'Anderson').

Example:

"filter": "LastName:R:Li"

True if the dimension is less than or equal to the value based on binary sort order.

Example:

"filter": "LastName:R::Levy"

True if the dimension and value are not equal.

Example:

"filter": "RoleName:N:Manager"

True if the dimension equals values val1 or val2. This filter expression uses the logical OR operator ( | ). You can compare the dimension value against more than two values. For example, to compare against three values, use the following syntax: dim1:EQ:val1|val2|val3.

Example:

"filter": "Lead

Status:EQ:Open|Contacted"

True if dimension dim1 equals value val1 and dimension dim2 equals value val2. This filter expression uses the logical AND operator ( , ). You can compare more than two dimensions. For example, to compare three dimensions, use the following syntax: dim1:EQ:val1,dim2:EQ:val2,dim3:EQ:val3.

Example:

"filter": "Lead

Status:EQ:Qualified,Rating:EQ:Hot"

SEE ALSO:

filter Transformation

flatten Transformation

The flatten transformation flattens hierarchical data. For example, you can flatten the Salesforce role hierarchy to implement row-level security on a dataset based on the role hierarchy.

When you configure the flatten transformation to flatten a hierarchy, you specify the field that contains every node in the hierarchy and the field that contains their corresponding parent based on the hierarchy. The flatten transformation generates one record for each hierarchy node, which we refer to as the “self ID.” Each record contains two generated columns that store the hierarchy for each self ID

131

Analytics Data Integration

Example:

Let’s look at an example. You want to create a dataset that contains all opportunities. For each opportunity, you want to include user and role information about the opportunity owner. Also, to implement row-level security based on the role hierarchy, each opportunity record must also contain a list of all ancestors above each opportunity owner based on the role hierarchy. To generate the list of ancestors, use the flatten transformation to flatten the role hierarchy.

You create the following dataflow definition file:

{

"Extract_Opportunity": {

"action": "sfdcDigest",

"parameters": {

132

Analytics

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "AccountId" },

{ "name": "OwnerId" }

]

}

"Extract_User": {

"action": "sfdcDigest",

"parameters": {

"object": "User",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "Department" },

{ "name": "UserRoleId" }

]

}

"Extract_UserRole": {

"action": "sfdcDigest",

"parameters": {

"object": "UserRole",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "ParentRoleId" }

]

}

"Flatten_UserRole": {

"action": "flatten",

"parameters": {

"source": "Extract_UserRole",

"self_field": "Id",

"parent_field": "ParentRoleId",

"multi_field": "Roles",

"path_field": "RolePath",

"include_self_field":false

}

"Augment_User_FlattenUserRole": {

"action": "augment",

"parameters": {

"left": "Extract_User",

"left_key": [ "UserRoleId" ],

"relationship": "Role",

"right": "Flatten_UserRole",

"right_key": [ "Id" ],

"right_select": [

133

Data Integration

Analytics Data Integration

"Id",

"Name",

"Roles",

"RolePath"

]

}

"Augment_Opportunity_UserWithRoles": {

"action": "augment",

"parameters": {

"left": "Extract_Opportunity",

"left_key": [ "OwnerId" ],

"right": "Augment_User_FlattenUserRole",

"relationship": "Owner",

"right_select": [

"Name",

"Department",

"Role.Id",

"Role.Name",

"Role.Roles",

"Role.RolePath"

"right_key": [ "Id" ]

}

"Register_OpportunityWithRoles_Dataset": {

"action": "sfdcRegister",

"parameters": {

"alias": "OppRoles",

"name": "OppRoles",

"source": "Augment_Opportunity_UserWithRoles",

"rowLevelSecurityFilter": "'Owner.Role.Roles' == \"$User.UserRoleId\" || 'OwnerId'

}

== \"$User.Id\""

}

To flatten the Salesforce role hierarchy, the flatten transformation uses the following input fields from the UserRole object.

Id identifies each node in the Salesforce role hierarchy.

ParentRoleId

ParentRoleId identifies the parent as defined in the role hierarchy.

After traversing through each parent-child relationship in the UserRole object, the flatten transformation generates one record for each role ID. Each record contains all ancestor roles for each role in the hierarchy. The flatten transformation generates two output columns—Roles and RolePath—to store all ancestor roles for each role.

IN THIS SECTION:

flatten Parameters

When you define a flatten transformation, you set the action attribute to flatten

and specify the parameters.

134

Analytics Data Integration

flatten Parameters

When you define a flatten transformation, you set the action attribute to flatten

and specify the parameters.

The following table describes the input parameters:

Parameter

include_self_id self_field parent_field multi_field path_field source

Required?

Yes

Value

Indicates whether to include the self ID node in the generated multi_field and path_field columns. Valid values are false

(default) and true

Name of the input field that identifies each node in the hierarchy.

Name of the input field that identifies the direct parent of each node in the hierarchy. For example, the Regional Manager 1 role is the parent of the Salesperson 1 role in a role hierarchy.

Name of the multi-value output field that contains a list of all ancestors in the hierarchy, in order from the lowest to the highest level. The flatten transformation creates this field and generates the list of ancestors for each node in the hierarchy. For example, for Salesperson 1 role, the hierarchy of ancestors is:

Sales

Manager 1, Regional Manager 1, Vice

President 1, CEO

A string representation of the multi-field field, separated by backslashes. This output field contains the hierarchical path of all ancestors in the hierarchy, in order from the lowest to the highest level. The flatten transformation creates this field and generates the ancestry path for each node in the hierarchy. For example, for a salesperson role in a role hierarchy, the value is:

Sales

Manager 1\Regional Manager 1\Vice President

1\CEO

Node in the dataflow definition file that contains the hierarchical data that you want to flatten. This node is the input source for this transformation and it must contain the input fields mapped to self_field and parent_field.

Note: By default, the multi_field and path_field fields are created as system fields, which aren’t visible in the user interface. To make the fields appear in the user interface, set the

IsSystemField

metadata attribute to

false

for each field in the flatten transformation. For more information about metadata attributes, see

Overriding Metadata Generated by a Transformation .

SEE ALSO:

flatten Transformation

135

Analytics Data Integration

sfdcDigest Transformation

When you upload the dataflow definition file, Wave Analytics validates access to Salesforce objects and fields based on the user profile of the user who uploads the file. If the user profile does not have read access to a field or object, the upload fails.

At run time, Wave Analytics runs the dataflow as the Integration User. Wave Analytics validates access to Salesforce objects and fields based on the user profile of the Integration User. For example, if the dataflow tries to extract data from a custom field on which the

Integration User does not have read access, the dataflow job fails.

Note: The Integration User is a preconfigured user that is created when Wave Analytics is enabled in your organization. If you or the Integration User need permission on a Salesforce object or field, ask the administrator to grant access.

For more information about preconfigured users in Wave Analytics, see the Wave Analytics Security Implementation Guide.

Example: Let’s look at an example. You would like to create a dataset that contains all opportunities from the Opportunity object.

You create the following dataflow definition.

{

"Extract_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "CloseDate" },

{ "name": "AccountId" },

{ "name": "OwnerId" },

{ "name": "OpportunitySupportTeamMembers__c" }

]

}

"Register_Opportunities_Dataset": {

"action": "sfdcRegister",

"parameters": {

"alias": "Opportunities",

"name": "Opportunities",

"source": "Extract_Opportunities"

}

Considerations When Using the sfdcDigest Transformation

•

Consider dataset storage limits when extracting data. For example, a dataset can contain a maximum of 5,000 fields, so be selective when choosing fields. See Wave Analytics Limits .

•

The sfdcDigest transformation runs a SOQL query to extract data from a Salesforce object, and so is subject to SOQL limits. If the query exceeds any of these limits, it may return no results or cause the dataflow job to fail. For example, The length of the SOQL query cannot exceed 20,000 characters. To reduce the SOQL query length, consider breaking up the extract into two or more

136

Analytics Data Integration sfdcDigest transformations and then use the augment transformation to combine the results. For example, you might create one sfdcDigest transformation to extract half of the fields and create another sfdcDigest transformation to extract the remaining fields.

See SOQL and SOSL Limits .

IN THIS SECTION:

Filtering Records Extracted from a Salesforce Object

Add a filter to the sfdcDigest transformation to extract a subset of all records from a Salesforce object. You can filter records to reduce the number of extracted and processed records, exclude records that contain irrelevant or sensitive data, and increase dataflow performance.

Overriding Salesforce Field Metadata

You can override the field metadata that the sfdcDigest transformation extracts from a Salesforce object to make the data appear differently in a dataset. For example, Wave Analytics can add a default value to records that have missing values for a field.

Unsupported Salesforce Objects and Fields in Wave

The sfdcDigest transformation can’t extract data from all Salesforce objects and fields. Consider these limitations before configuring the extraction of Salesforce objects.

sfdcDigest Parameters

When you define an sfdcDigest transformation, you set the action attribute to sfdcDigest

and specify the parameters for the object and fields that you want to extract. Optionally, you can also specify parameters to filter the records extracted from the Salesforce object.

Filtering Records Extracted from a Salesforce Object

A filter consists of one or more filter conditions, where each filter condition compares a field value to a value. For example,

Amount

>= 1000000

. You can also apply SOQL functions on the field value in a filter condition, like

CALENDAR_YEAR(CreatedDate)

= 2011

. You can add multiple filter conditions using logical operators AND, OR, and NOT. You can also use a backslash (\) to escape double quotes included in strings.

The sfdcDigest transformation extracts all records for which the filter is true. If you configured the sfdcDigest transformation for incremental extraction, the filter applies to data extracted during the incremental run only—Wave Analytics doesn't apply the filter to records that were previously loaded into the dataset. If you add an invalid filter, the dataflow fails at run time.

For each instance of sfdcDigest, you can use one of the following types of filters:

•

Structured filter

•

Advanced filter

Tip: Are you trying to decide whether to use a filter in the sfdcDigest transformation or use a filter transformation? Use a filter transformation to filter records at any point in the dataflow. For example, you can add it after the dataflow joins two datasets.

However, to reduce the number of rows processed in the dataflow and optimize dataflow performance, add the filter closest to the point at which records are extracted—when possible, add the filter in the sfdcDigest transformation.

IN THIS SECTION:

Structured Filter in sfdcDigest Transformation

You define a structured filter using JSON syntax.

137

Analytics Data Integration

Advanced Filter in sfdcDigest Transformation

You define an advanced filter using a Salesforce Object Query Language (SOQL) WHERE clause expression. Use an advanced filter only if you are familiar with SOQL.

SEE ALSO:

Structured Filter in sfdcDigest Transformation

You define a structured filter using JSON syntax.

A structured filter uses the following JSON syntax for each filter condition.

{

"field": "<field name>",

"operator": "<operator>",

"value": "<value>"|"["<value 1>", "<value 2>"]",

"isQuoted": true|false}

The value can be a number, date, string, list of strings, or date literal . Wave Analytics automatically quotes strings unless you set isQuoted to true, which indicates that the string is already quoted.

You can use one of the following operators in a filter condition.

Operator

Comment

Filter condition is true if the value in the field equals the specified value. String comparisons using the equals operator are case-insensitive.

Example:

"filterConditions": [

{

"field": "OwnerId",

"operator": "=",

"value": "a07B00000012HYu"

}

]

Filter condition is true if the value in the field does not equal the specified value.

Example (using backslashes to escape double quotes in a string value):

"filterConditions": [

{

"field": "Nickname__c",

"operator": "!=",

"value": "\"Sammy\""

}

]

138

Analytics

Operator

139

Data Integration

Comment

Filter condition is true if the value in the field is greater than the specified value.

Example:

"filterConditions": [

{

"field": "Amount",

"operator": ">",

"value": "100000"

}

]

Filter condition is true if the value in the field is less than the specified value.

Example (using a date literal):

"filterConditions": [

{

"field": "CloseDate",

"operator": "<",

"value": "THIS_MONTH",

"isQuoted": false

}

]

Filter condition is true if the value in the field is greater than or equal to the specified value.

Example:

"filterConditions": [

{

"field": "Amount",

"operator": ">=",

"value": "100000"

}

]

Filter condition is true if the value in the field is less than or equal to the specified value.

Example (using a SOQL function):

"filterConditions": [

{

"field": "CALENDAR_YEAR (CreatedDate)",

"operator": "<=",

"value": "2015",

"isQuoted": true

Analytics

Operator

NOT IN

140

Data Integration

Comment

}

]

Filter condition is true if the value in the field matches the specified value. The LIKE operator is similar to the LIKE operator in SQL; it provides a mechanism for matching partial text strings and supports wildcards.

•

The % and _ wildcards are supported for the LIKE operator.

•

The % wildcard matches zero or more characters.

•

The _ wildcard matches exactly one character.

•

The LIKE operator is supported for string fields only.

•

The LIKE operator performs a case-insensitive match.

•

The LIKE operator supports escaping of special characters % or _. Use a backslash (\) to escape special characters.

Example:

"filterConditions": [

{

"field": "FirstName",

"operator": "LIKE",

"value": "Chris%"

}

]

Filter condition is true if the value in the field equals any one of the values in the specified list. You can specify a quoted or non-quoted list of values. If the list is quoted, set isQuoted

to true.

Example:

"filterConditions": [

{

"field": "StageName",

"operator": "IN",

"value": ["Closed Won", "Closed Lost"]

}

]

Filter condition is true if the value in the field does not equal any of the values in the specified list.

Example:

"filterConditions": [

{

"field": "BillingState",

"operator": "NOT IN",

"value": ["California", "New York"]

Analytics

Operator

INCLUDES

EXCLUDES

Data Integration

Comment

}

]

For picklist or multi-select picklist fields only. Filter condition is true if the value in the picklist field includes the specified value.

Example:

"filterConditions": [

{

"field": "BillingState",

"operator": "INCLUDES",

"value": ["California"]

}

]

For picklist or multi-select picklist fields only. Filter condition is true if the value in the picklist field excludes the specified value.

Example:

"filterConditions": [

{

"field": "BillingState",

"operator": "EXCLUDES",

"value": ["California", "New York"]

}

]

Let’s look at a few examples of structured filters.

Example: Let’s look at an example with a basic structured filter. To perform pipeline analysis on opportunities in fiscal quarter 2 of fiscal year 2015, you create this dataflow definition file to create the relevant dataset.

{

"Extract_Filtered_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "AccountId" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "CloseDate" },

{ "name": "OwnerId" },

{ "name": "FiscalYear" },

{ "name": "FiscalQuarter" },

{ "name": "SystemModstamp" }

141

Analytics Data Integration

"filterConditions": [

{

"field": "FiscalYear",

"operator": "=",

"value": "2015"

{

"field": "FiscalQuarter",

"operator": "=",

"value": "2"

}

]

}

"Register_Opportunities_Dataset": {

"action": "sfdcRegister",

"parameters": {

"alias": "Opportunities_2015Q2",

"name": "Opportunities_2015Q2",

"source": "Extract_Filtered_Opportunities"

}

Note: If you do not specify a logical operator for multiple filter conditions—as is the case in this example—Wave Analytics applies AND between the conditions.

Example: Let's look at an example of a structured filter with a logical operator. To help forecast expected revenue, you create this dataflow to view all opportunities that have either closed or have greater than 90% probability of closing.

{

"Extract_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "AccountId" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "CloseDate" },

{ "name": "OwnerId" },

{ "name": "Probability" },

{ "name": "FiscalYear" },

{ "name": "FiscalQuarter" }

"filterConditions": [

{

"operator": "OR",

"conditions": [

{

"field": "StageName",

142

Analytics Data Integration

}

"operator": "=",

"value": "Closed Won"

{

}

]

}

]

}

"Register_Opportunities_Dataset": {

"action": "sfdcRegister",

"parameters": {

"alias": "OpportunitiesExpectedToWin",

"name": "OpportunitiesExpectedToWin",

"source": "Extract_Opportunities"

}

"field": "Probability",

"operator": ">=",

"value": "90"

Example: Finally, let's look at an example of a structured filter with nested logical operators. You create this dataflow to view all opportunities that closed in the current fiscal quarter and are owned by either one of your two direct reports.

{

"Extract_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "AccountId" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "CloseDate" },

{ "name": "OwnerId" },

{ "name": "FiscalYear" },

{ "name": "FiscalQuarter" }

"filterConditions": [

{

"operator": "AND",

"conditions": [

{

"field": "CloseDate",

"operator": "=",

"value": "THIS_FISCAL_QUARTER",

"isQuoted": false

{

"operator": "OR",

143

Analytics Data Integration

}

"conditions": [

{

"field": "OwnerId",

"operator": "=",

"value": "00540000000HfUz"

{

"field": "OwnerId",

"operator": "=",

"value": "00540000000HfV4"

}

]

}

]

}

]

}

"Register_Opportunities_Dataset": {

"action": "sfdcRegister",

"parameters": {

"alias": "DirectReport_Opportunities",

"name": "DirectReport_Opportunities",

"source": "Extract_Opportunities"

}

Advanced Filter in sfdcDigest Transformation

You define an advanced filter using a Salesforce Object Query Language (SOQL) WHERE clause expression. Use an advanced filter only if you are familiar with SOQL.

Example: Let’s look at an example of an advanced filter. You want to extract only opportunity records that are owned by a specific user and that have either high value or a high probability of closing. You create this dataflow.

{

"Extract_Filtered_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "AccountId" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "CloseDate" },

{ "name": "Probability" },

{ "name": "OwnerId" }

"complexFilterConditions": "OwnerId = '00540000000HfUz' AND (Amount > 100000

OR Probability > 75)"

144

Analytics Data Integration

}

"Register_Opportunities_Dataset": {

"action": "sfdcRegister",

"parameters": {

"alias": "FilteredOpportunities",

"name": "FilteredOpportunities",

"source": "Extract_Filtered_Opportunities"

}

Overriding Salesforce Field Metadata

You can add the following field parameters to the sfdcDigest transformation to override the field metadata:

•

defaultValue

•

type

•

fiscalMonthOffset

•

isYearEndFiscalYear

•

firstDayOfWeek

•

isMultiValue

•

multiValueSeparator

•

precision

•

scale

For a description of each of these field parameters, see

Field Parameters

. For information about using metadata attributes to configure dates, see Date Handling in Datasets.

Example:

Let’s look at an example. You would like to override metadata extracted from the Opportunity object.

You add the bold text in the following dataflow definition to override field metadata from the Opportunity object.

{

"Extract_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{

"name": "Amount",

"defaultValue":0

"precision":18

"scale":2

{ "name": "StageName" },

{

145

Analytics Data Integration

}

"name": "CloseDate",

"fiscalMonthOffset":9,

"firstDayOfWeek":2,

"isYearEndFiscalYear":true

{ "name":"AccountId" },

{ "name":"OwnerId" },

{

"name": "OpportunitySupportTeamMembers__c",

"type":"Text",

"isMultiValue":true,

"multiValueSeparator":","

"precision":255

}

]

}

"Register_Opportunities_Dataset":{

"action":"sfdcRegister",

"parameters":{

"alias":"Opportunities",

"name":"Opportunities",

"source":"Extract_Opportunities"

}

SEE ALSO:

Unsupported Salesforce Objects and Fields in Wave

The sfdcDigest transformation can’t extract data from all Salesforce objects and fields. Consider these limitations before configuring the extraction of Salesforce objects.

For information about all Salesforce objects and fields, see the Object Reference for Salesforce and Force.com

Unsupported Objects

The sfdcDigest transformation can’t extract data from these Salesforce objects.

•

AuthProvider

•

BrandTemplate

•

ChatterConversation

•

ChatterConversationMember

•

ChatterMessage

•

ConnectedApplication

•

ContentFolderLink

•

ContentWorkspace

•

ContentWorkspaceDoc

146

Analytics

•

CorsWhitelistEntry

•

CustomNotDeployed__OwnerSharingRule

•

EmailDomainKey

•

EmailServicesAddress

•

EmailServicesFunction

•

EmailStreamMessage

•

EmailStreamShareOption

•

EmailStreamShareOptionHistory

•

EmailStreamShareOptionOwnerSharingRule

•

EmailStreamShareOptionShare

•

EmailTemplate

•

EventType

•

EventTypeParameter

•

EnvironmentHub

•

EnvironmentHubInvitation

•

EnvironmentHubMemberRel

•

FeedPollChoice

•

FeedPollVote

•

KnowledgeArticleVersion

•

LoginGeo

•

LoginHistory

•

NetworkActivityAudit

•

NewsArticleFeed

•

NetworkModeration

•

OrganizationProperty

•

OrgWideEmailAddress

•

OutboundField

•

PackageLicense

•

PackageMemberOwnerSharingRule

•

PartnerNetworkSyncLog

•

PermissionSet

•

PermissionSetLicense

•

Profile

•

ReputationLevel

•

ReputationLevelLocalization

•

ReputationPointsRule

•

SearchPromotionRule

•

SelfServiceUser

•

SessionPermSetAssignment

•

SetupAssistantAnswer

147

Data Integration

Analytics Data Integration

•

SetupAssistantProgress

•

SsoUserMapping

•

TenantSecret

•

Territory2ModelHistory

•

TwoFactorInfo

•

UserLogin

•

UserPackageLicense

•

UserProvAccount

•

UserProvAccountStaging

•

UserProvisioningConfig

•

UserProvisioningLog

•

UserProvisioningRequest

•

UserProvisioningRequestOwnerSharingRule

•

UserProvisioningRequestShare

•

UserProvMockTarget

•

UserRecordAccess

•

VerificationHistory

•

WebLink

•

WebLinkLocalization

The sfdcDigest transformation cannot extract data from external objects created in Salesforce. External objects are similar to custom objects, except that they map to data located outside Salesforce.

If you include an unsupported or inaccessible object in the sfdcDigest transformation, the dataflow fails at run time with an error message.

Unsupported Fields

The sfdcDigest transformation cannot extract data from these fields:

Object

Account

ActionPlanItem

AuthSession

CaseArticle

Contact

ContentDocument

ContentFolderLink

CustomPerson__p

Unsupported Fields

CleanStatus

ItemId

•

LoginGeoId

•

LoginHistoryId

KnowledgeArticleId

•

CanAllowPortalSelfReg

•

CleanStatus

ParentId

ParentEntityId

Title

148

Analytics

Object

DocumentAttachmentMap

EmailMessage

EmailRoutingAddress

EnvironmentHubMember

ExternalEventMapping

InstalledMobileApp

Lead

KnowledgeArticle

KnowledgeArticleVersion

Network

Organization

PermissionSet

PermissionSetLicense

Profile

ThirdPartyAccountLink

149

Unsupported Fields

ParentId

ActivityId

EmailServicesAddressId

EnvironmentHubId

EventId

ConnectedApplicationId

CleanStatus

MasterLanguage

•

IsOutOfDate

•

TranslationCompletedDate

•

TranslationExportedDate

•

TranslationImportedDate

•

CaseCommentEmailTemplateId

•

ChangePasswordEmailTemplateId

•

ForgotPasswordEmailTemplateId

•

WelcomeEmailTemplateId

•

SelfServiceEmailUserOnCaseCreationTemplateId

•

SelfServiceNewCommentTemplateId

•

SelfServiceNewPassTemplateId

•

SelfServiceNewUserTemplateId

•

WebToCaseAssignedEmailTemplateId

•

WebToCaseCreatedEmailTemplateId

•

WebToCaseEmailTemplateId

•

WebToLeadEmailTemplateId

•

PermissionsEditEvent

•

PermissionsEditTask

•

MaximumPermissionsEditEvent

•

MaximumPermissionsEditTask

•

PermissionsEditEvent

•

PermissionsEditTask

SsoProviderId

Data Integration

Analytics Data Integration

Object

User

WorkBadge

WorkBadgeDefinition

Unsupported Fields

•

LastPasswordChangeDate

•

UserPreferencesEnableVoicePilot

RewardId

RewardFundId

If you include a field with an unsupported field in the sfdcDigest transformation, the dataflow ignores the field.

In addition, Salesforce recommends that you do not extract data from the MayEdit field of the Account object. Extracting data from this field significantly decreases performance and can cause the dataflow to fail.

Unsupported Field Types

The sfdcDigest transformation can’t extract data from fields with these field types.

•

base64

•

composite (like address and location)

•

data category group reference

•

encrypted string

If you include a field with an unsupported field type in the sfdcDigest transformation, the dataflow ignores the field.

SEE ALSO:

For more information, see Date Handling in

sfdcDigest Parameters

When you define an sfdcDigest transformation, you set the action attribute to sfdcDigest

and specify the parameters for the object and fields that you want to extract. Optionally, you can also specify parameters to filter the records extracted from the Salesforce object.

You can specify parameters in the following sections of the sfdcDigest node: parameters, fields, and filterConditions.

Parameters

The following table describes the parameters in the parameters

section.

Parameter

object

Required?

Yes incremental No

Value

API name of the Salesforce object from which you want to extract data. This object is the input source for this transformation.

The sfdcDigest transformation doesn’t support extraction from all Salesforce objects.

Performs an incremental extraction, which extracts only changes to the Salesforce

150

Analytics Data Integration

Parameter

fullRefreshToken fields filterConditions complexFilterConditions

Required?

Yes

Value

object since the last dataflow run. Valid values: true or false.

Note:

Incremental extraction is available if you have enabled replication.

Performs a one-time full extraction to synchronize the data in the dataset with data in the Salesforce object. Specify any value for this parameter.

After the full extraction, the dataflow performs an incremental extraction each time thereafter even though the fullRefreshToken

parameter is included in the dataflow definition. To run a full extraction again, change the value of the fullRefreshToken

parameter to a different value.

Note: Incremental extraction is available if you have enabled replication.

An array of names of all fields from which you want to extract data from the specified

Salesforce object. The sfdcDigest transformation doesn’t support extraction from all field types.

See Field Attributes.

A filter that restricts the records extracted from the specified Salesforce object. The sfdcDigest transformation extracts all records from the Salesforce object for which the filter is true. You can specify a structured or advanced filter.

See Filter Conditions Parameters.

For advanced filters only. A SOQL WHERE clause used to filter records extracted from the specified Salesforce object.

Field Attributes

The following table describes the attributes in the fields

section. It also describes optional attributes that you can provide to override the field metadata. You can override the metadata that the sfdcDigest transformation extracts from a Salesforce object to make the data

151

Analytics Data Integration appear differently in a dataset. For example, Wave Analytics can add a default value to records that have missing values for a field. If you don’t override the values, Wave Analytics gets the values from Salesforce.

Attribute

name defaultValue

Required?

Yes

Value

API name of the field in the Salesforce object that you want to include in the dataset. You can specify multiple fields.

For text and numeric fields that can be null.

Default value that replaces a null value for the specified field.

type No

Wave Analytics field type associated with the specified field. Valid types are Text,

Numeric, or Date. Any value, including numeric values, can be Text. For example, by default, fiscal quarter from Salesforce objects is Number. However, you can change it to Text. Specify a type to override the type determined by Wave Analytics.

Example:

"type": "Text" fiscalMonthOffset No

For date fields only. The difference, in months, between the first month of the fiscal year and the first month of the calendar year (January). For example, if the fiscal year starts in January, the offset is 0. If the fiscal year starts in October, the offset is

Example:

"fiscalMonthOffset": 9

Note: This attribute also controls whether Wave Analytics generates fiscal date fields. To generate fiscal date fields, set fiscalMonthOffset

to a value other than 0.

For more information, see Date Handling in

isYearEndFiscalYear No

For date fields only. Indicates whether the fiscal year is the year in which the fiscal year ends or begins. Because the fiscal year can start in one calendar year and end in

152

Analytics

Attribute

firstDayOfWeek No isMultiValue multiValueSeparator

Required?

Data Integration

Value

another, you must specify which year to use for the fiscal year.

•

If true, then the fiscal year is the year in which the fiscal year ends. The default is true.

•

If false, then the fiscal year is the year in which the fiscal year begins.

Example:

"isYearEndFiscalYear": true

This field is relevant only when fiscalMonthOffset

is greater than

For more information, see Date Handling in

For date fields only. The first day of the week for the calendar year and, if applicable, fiscal year. Use 0 to set the first day to be Sunday,

1 to set the first day to be Monday, and so on. Use -1 to set the first day to be on

January 1. The default is -1.

Example:

"firstDayOfWeek": 0

For text fields only. Indicates whether the specified field has multiple values.

Example:

"isMultiValue": false

For text fields only. Character used to separate multiple values in the specified field when isMultiValue equals true. This value defaults to a semicolon (;) if you do not specify a value and isMultiValue equals true. Set to null when isMultiValue equals false.

Example:

"multiValueSeparator": ";"

153

Analytics

Attribute

precision scale

Required?

Data Integration

Value

The maximum number of digits in a numeric value, or the length of a text value. For numeric values: Includes all numbers to the left and to the right of the decimal point

(but excludes the decimal point character).

Value must be between 1 and 18, inclusive.

For text values: Value defaults to 255 characters, and must be between 1 and

32,000 characters, inclusive.

Example:

"precision": 10

The number of digits to the right of the decimal point in a numeric value. Must be less than the precision value. Value must be between 0 and 17 characters, inclusive.

Example:

"scale": 2

Filter Conditions Parameters

The following table describes the structured filter parameters in the filterConditions

section. These parameters do not apply to advanced filters.

Parameter

field

Required?

Value

The field in the Salesforce object on which you want to apply a filter condition. Each filter condition in a structured filter uses the following syntax:

{

"field": "<field name>",

"operator": "<operator>",

"value": "<value>",

"isQuoted": true|false} operator No

The purpose depends on the context.

•

operator

can be used as a comparison operator–like =, <, and

IN–that compares the field value against a constant value.

154

Analytics

Parameter Required?

value isQuoted

155

Data Integration

Value

•

operator

can also be used as a logical operator (AND, OR, or NOT) that links multiple filter conditions together.

In the example below, the bold operator

is the logical operator. The other instances of operator

are comparison operators.

"filterConditions": [

{

"operator": "OR",

"conditions": [

{

"field":

"StageName",

"operator": "=",

"value": "Closed

Won"

{

"field":

"Probability",

"operator":

">=",

"value": "90"

}

]

}

]

The value used in a filter condition.

Indicates whether you quoted the string value in a filter condition.

Example with quoted values:

"filterConditions": [

{

"field": "StageName",

"operator": "IN",

"value": "('Closed Won',

'Closed Lost')",

"isQuoted": true

}

]

Example with non-quoted values:

"filterConditions": [

{

Analytics Data Integration

Parameter

conditions

Required?

Value

"value": ["Closed Won",

"Closed Lost"],

"isQuoted": false

}

]

"field": "StageName",

"operator": "IN",

If you don’t include isQuoted for a filter on a string value, Wave Analytics assumes that the string value is not quoted and adds the quotes for you.

Use to specify a logical operator to link multiple filter conditions together.

SEE ALSO:

Filtering Records Extracted from a Salesforce Object

sfdcRegister Transformation

The sfdcRegister transformation registers a dataset to make it available for queries. Users cannot view or run queries against unregistered datasets.

You don’t need to register all datasets. For example, you don’t need to register an intermediate dataset that is used to build another dataset and does not need to be queried. In addition, you don’t need to register datasets that are created when you upload external data because Wave Analytics automatically registers these datasets for you.

Carefully choose which datasets to register because:

•

The total number of rows in all registered datasets cannot exceed 250 million per platform license.

•

Users that have access to registered datasets can query their data. Although, you can apply row-level security on a dataset to restrict access to records.

Example: Let’s look at an example. You create a dataflow that extracts opportunities from the Opportunity object. To register the dataset, name it "Opportunities," and apply row-level security on it, you add the sfdcRegister transformation as shown in the following dataflow definition file.

{

"Extract_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "CloseDate" },

156

Analytics Data Integration

}

{ "name": "AccountId" },

{ "name": "OwnerId" }

]

}

"Register_Oppportunities_Dataset": {

"action": "sfdcRegister",

"parameters": {

"alias": "Opportunities",

"name": "Opportunities",

"source": "Extract_Opportunities",

"rowLevelSecurityFilter": "'OwnerId' == \"$User.Id\""

}

IN THIS SECTION:

sfdcRegister Parameters

When you define an sfdcRegister transformation, you set the action attribute to sfdcRegister

and specify the parameters.

sfdcRegister Parameters

When you define an sfdcRegister transformation, you set the action attribute to sfdcRegister

and specify the parameters.

The following table describes the input parameters:

Parameter

alias name source

Required?

Yes

Value

API name of the registered dataset. This name can contain only underscores and alphanumeric characters, and must be unique among other dataset aliases in your organization. It must begin with a letter, not include spaces, not end with an underscore, and not contain two consecutive underscores. It also cannot exceed 80 characters.

Display name of the registered dataset. The name cannot exceed 80 characters.

Note:

To change the name after you create the dataset, you must edit the dataset.

Node in the dataflow definition file that identifies the dataset that you want to register. This is the input source for this transformation.

157

Analytics

Parameter

rowLevelSecurityFilter

Required?

Data Integration

Value

The predicate used to apply row-level security on the dataset when the dataset is first created.

Example: "rowLevelSecurityFilter":

"'OwnerId' == "$User.Id""

Note:

To change the predicate after you create the dataset, you must edit the dataset.

When entering the predicate in the

After the dataset is created, Wave Analytics ignores it's security predicate setting in the dataflow. To change the security predicate for an existing dataset, edit the dataset in the user interface.

SEE ALSO:

sfdcRegister Transformation

sliceDataset Transformation

The transformation includes the mode

parameter, which lets you choose between two modes: drop

or select

. Set it to drop

remove the fields you specify. Set it to select

to keep the fields you specify.

Example: Here’s an example. Let’s say your dataflow contains an sfdcDigest node,

ExtractOpportunities

, that extracts all fields from the Opportunity object. However, some of your users don’t want Id fields in their Opportunities dataset. You resolve this by adding a sliceDataset node to your dataflow, specifying the source node, the fields you want to remove from that node, and drop

mode.

{

"RemoveFieldsFromExtractOpportunities": {

"action": "sliceDataset",

"parameters": {

"source": "ExtractOpportunities",

"mode": "drop",

"fields": [

{

"name": "Id"

{

"name": "ProductId"

158

Analytics Data Integration

}

]

}

IN THIS SECTION:

sliceDataset Parameters

When you define a sliceDataset transformation, you set the action attribute to sliceDataset

and specify the parameters.

sliceDataset Parameters

When you define a sliceDataset transformation, you set the action attribute to sliceDataset

and specify the parameters.

The following table describes the input parameters:

Parameter

source mode

Required?

Yes

Value

Node in the dataflow definition file that identifies the dataset that you want to slice.

This is the input source for this transformation.

The action you want to take on the specified fields. Set it to drop

to remove the fields.

Set it to select

to keep the fields.

fields Yes

An array of names of the fields that you want to drop or select from the specified source node.

update Transformation

When you create the transformation, you specify the keys that are used to match records between the two datasets. To dictate which field in the lookup dataset updates the field in the source dataset, you also map the corresponding fields from both datasets.

Example: Let’s look at an example. You have an existing Accounts dataset that contains account information—Id, Name, and

AnnualRevenue. Unfortunately, some of the account names in the dataset are now incorrect because of a series of mergers and acquisitions. To quickly update the account names in the dataset, you perform the following tasks.

1. Create a .csv file that contains the new account names and associated account IDs for accounts that have name changes.

2. Upload the .csv file to create a dataset called UpdatedAccountNames.

3. Create a dataflow definition file to update account names in the Accounts dataset by looking up the new account names in the UpdatedAccountNames dataset.

159

Analytics Data Integration

You create the following dataflow definition file.

{

"Extract_AccountDetails": {

"action": "sfdcDigest",

"parameters": {

"object": "Account",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "AnnualRevenue" }

]

}

"Extract_UpdatedAccountNames": {

"action": "edgemart",

"parameters": { "alias": "UpdatedAccountNames" }

"Update_AccountRecords": {

"action": "update",

"parameters": {

"left": "Extract_AccountDetails",

"right": "Extract_UpdatedAccountNames",

"left_key": [ "Id" ],

"right_key": [ "AccountID" ],

"update_columns": { "Name": "NewAccountName" }

}

"Register_UpdatedAccountRecords": {

"action": "sfdcRegister",

"parameters": {

"alias": "Accounts",

"name": "Accounts",

"source": "Update_AccountRecords"

}

Example: Let’s look at another example, where a composite key is used to match records between both datasets. In this case, you match records using the account ID and account name fields.

You create the following dataflow definition file.

{

"Extract_AccountDetails": {

160

Analytics Data Integration

}

"action": "sfdcDigest",

"parameters": {

"object": "Account",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "AnnualRevenue" }

]

}

"Extract_UpdatedAccountNames": {

"action": "edgemart",

"parameters": { "alias": "UpdatedAccountNames" }

"Update_AccountRecords": {

"action": "update",

"parameters": {

"left": "Extract_AccountDetails",

}

"Register_UpdatedAccountRecords": {

"action": "sfdcRegister",

"parameters": {

"alias": "Accounts",

"name": "Accounts",

"source": "Update_AccountRecords"

}

"right": "Extract_UpdatedAccountNames",

"left_key": ["Id","Name"],

"right_key": ["AccountId","NewAccountName"],

"update_columns": {

"Name": "NewAccountName",

"CreatedDate":"NewCreatedDate",

"AnnualRevenue":"NewAnnualRevenue"

IN THIS SECTION:

update Parameters

When you define an update transformation, you set the action attribute to update

and specify the parameters.

update Parameters

When you define an update transformation, you set the action attribute to update

and specify the parameters.

The following table describes the input parameters.

Parameter

left

Required?

Yes

Value

Node in the dataflow definition file that identifies the dataset that contains the records that you want to update.

161

Analytics

Parameter

right left_key right_key update_columns

Yes

Required?

Yes

Data Integration

Value

Node in the dataflow definition file that identifies the lookup dataset that contains the new values.

Key column in the left dataset used to match records in the other dataset. If you use a composite key, the left and right keys must have the same number of columns in the same order. For an example, see

update

Transformation on page 159.

Key column in the right dataset used to match records in the other dataset. If you use a composite key, the left and right keys must have the same number of columns in the same order.

An array of corresponding columns between the left and right datasets. Use the following syntax:

"update_columns": {

"LeftColumn1":

"RightColumn1",

"LeftColumn2":

"RightColumn2",...

"LeftColumnN":

"RightColumnN" }

. The value from right column replaces the value from the corresponding left column. The field types of the left and right column must match.

Note: If you specify a column name that does not exist, the dataflow fails.

If you do not specify this parameter, the transformation updates the left dataset by matching all columns in the right dataset with those in the left. In this case, the right column names must match exactly with the left column names. Otherwise, an error might occur.

SEE ALSO:

update Transformation

162

Analytics Data Integration

Overriding Metadata Generated by a Transformation

Optionally, you can override the metadata that is generated by a transformation. You can override object and field attributes. For example, you can change a field name that is extracted from a

Salesforce object so that it appears differently in the dataset. To override the metadata, add the overrides to the Schema section of the transformation in the dataflow definition file.

In the Schema section, you can override the metadata attributes for one object only.

The Schema section in this sample sfdcDigest transformation contains metadata overrides:

EDITIONS

Available in: Salesforce

Classic and Lightning

Experience

Available for an additional cost in: Enterprise,

Performance, and

Unlimited Editions

"Extract_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Name" },

{ "name": "Amount" }

]

"schema": {

"objects": [

{

"label":"Sales Opportunities",

"fields": [

{

"name": "Amount",

"label": "Opportunity Amount"

}

]

}

]

}

Object Attributes

You can override the following object attributes.

Object Attribute

label

Type

String description

String

Description

The display name for the object. Can be up to 40 characters.

Example:

"label": "Sales Data"

The description of the object. Must be less than 1,000 characters.

Add a description to annotate an object in the dataflow definition file. This description is not visible to users in the Wave Analytics user interface.

163

Analytics

Object Attribute

fields

Type

Array

Data Integration

Description

Example:

"description": "The SalesData object tracks basic sales data."

The array of fields for this object.

Field Attributes

You can override attributes of each specified dataset field.

Field Attribute

name

Type

String

Description

Name of the field in the dataset. Identifies the field that you want to override.

Examples:

"name": "Amount"

"name": "Role.Name" label

String

The display name for the field. Can be up to 255 characters.

Example:

"label": "Opportunity Amount" description

String

The description of the field. Must be less than 1,000 characters.

Add a description to annotate a field in the dataflow definition file. This description is not visible to users in the Wave Analytics user interface.

Example:

"description": "The Amount field contains the opportunity amount." isSystemField

Boolean format

String

Indicates whether this field is a system field to be excluded from query results.

Example:

"isSystemField": false

The display format of the numeric value.

Examples:

"format": "$#,##0.00" (Numeric)

For more information about valid formats, see Numeric Formats

Numeric Formats

An example of a typical numeric value is $1,000,000.99, which is represented as $#,##0.00. You are required to specify the precision and scale of the number. The format is specified by using the following symbols:

164

Analytics

Symbol

Data Integration

Meaning

One digit

Zero or 1 digit

This symbol is the default decimal separator. Use the decimalSeparator

field to set the decimal separator to a different symbol.

Minus sign

Grouping separator

Currency sign

Note: The format for numeric values isn’t used in data ingestion. It is used only to specify how numeric values are formatted when displayed in the UI. Also, you can’t override date formats.

Example:

Let’s consider an example where you want to override the following object and field attributes that the sfdcDigest transformation extracts from the Opportunity object.

Object/Field

Opportunity object

Attribute Changes

•

Change the object label to "Sales Opportunities"

•

Add an object description

Id field

•

Change the field label to "Opportunity Id"

•

Hide the field from queries

Amount field

•

Change the field label to "Opportunity Amount"

•

Change the format to $#,##0.00

CloseDate field

•

Change the field label to "Closing Date"

To override the attributes, you add the Schema section with the override values to sfdcDigest in the dataflow definition file.

{

"Extract_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "CloseDate" },

{ "name": "AccountId" },

{ "name": "OwnerId" }

165

Analytics

]

"schema": {

"objects": [

{

"label":"Sales Opportunities",

"description": "These are all sales opportunities.",

"fields": [

{

"name": "Id",

"label": "Opportunity Id",

"isSystemField": true

{

"name": "Amount",

"label": "Opportunity Amount",

"format": "$#,##0.00"

{

"name": "CloseDate",

"label": "Closing Date"

}

]

}

]

}

"Register_Dataset_Opportunities": {

"action": "sfdcRegister",

"parameters": {

"source": "Extract_Opportunities",

"alias": "Opportunities",

"name": "Opportunities"

}

Data Integration

166

Analytics Data Integration

Create a Dataset with the Dataset Builder

You can use the dataset builder to create a single dataset based on data from one or more Salesforce objects. The dataset builder generates and appends the associated JSON to the dataflow definition file. The dataset is created the next time the dataflow runs. The data in the dataset refreshes each time the dataflow runs. You can also edit the dataflow definition file to add transformations that manipulate the dataset.

1. On the home page or on an app page, click Create > Dataset.

2. Click Salesforce.

The dataset builder opens on the Dataset Builder tab.

EDITIONS

Available in: Salesforce

Classic and Lightning

Experience

Available for an additional cost in: Enterprise,

Performance, and

Unlimited Editions

USER PERMISSIONS

To access the dataset builder:

•

“Edit Wave Analytics

Dataflows”

3. Select the root object.

The root object is the lowest level child object that you can add to the canvas. After you select the root object, you can add only parent objects of the root object—you can’t add it’s children objects. To change the root object, refresh the page and start over.

4. Hover over the root object, and then click

The Select Fields dialog box appears. By default, the Fields tab appears and shows all available object fields from which you can extract data.

167

Analytics Data Integration

Note:

You can view this dialog box for any object included in the canvas.

5. In the Fields tab, select the fields from which you want to extract data.

To locate fields more quickly, you can search for them or sort them by name or type.

Important:

You must select at least one field for each object that you add to the canvas. If you add an object and don’t add any of it’s fields, the dataflow fails at run time.

6. In the Relationships tab, click Join to add the related objects to the canvas.

When you add a related object, the related object appears in the canvas.

168

Analytics Data Integration

7. To remove a related object, click Delete.

Warning:

When you delete a related object, you also delete all objects that descend from the related object in the diagram.

For example, if you delete Account shown below, you delete the branch that contains Account and User.

8. For each related object, select the fields from which you want to extract data.

9. To move the entire diagram, select a white space in the canvas and drag it.

You might need to move the diagram to view a different section of the diagram.

10. To view all objects included in the canvas, click

The Selected Objects dialog box shows a tree structure of all objects included in the canvas. The root object appears at the top of the tree.

If you select one of the objects, the dataset builder focuses on the object by placing the object in the center of the canvas.

11. To view the associated JSON, click

When you create the dataset, the dataset builder appends the JSON to the dataflow definition file.

169

Analytics Data Integration

12. Click Create Dataset.

13. Enter the name of the dataset, and select the app that will contain the dataset if it’s not already selected.

Note:

If you enter a dataset name that is already used, when you create the dataset, the dataset builder appends a number to the dataset name. For example, if you entered MyOpportunities, the dataset builder creates MyOpportunities1. The dataset name cannot exceed 80 characters.

14. Click Create.

The dataset builder appends the underlying JSON to the dataflow definition file. The dataset is created the next time the dataflow runs. You can manually run the dataflow to immediately create the dataset.

Optimize Your Dataflows with Replication

Without replication, a dataflow performs a separate extract each time it needs data from a Salesforce object. Let’s take a look at an example. Imagine your organization has three dataflows, extracting data from Salesforce objects as follows:

Accounts Contacts Opportunities Campaigns Leads Cases Users

Default

Dataflow

Sales Wave

App Dataflow

Service Wave

App Dataflow

Every time these dataflows run, they have to extract all this Salesforce data. And the more data there is, the longer the dataflow takes to run. What’s more, the dataflows are duplicating effort by performing separate extracts from the same object. For example, Accounts data is extracted by all three dataflows.

With replication, all of these extracts are performed as a separate process, which you can schedule to take place before your dataflows run.

Accounts Contacts Opportunities Campaigns Leads Cases Users

Replication

This replicated data is then available to all your dataflows, which run faster because they no longer have to extract any data—just load and transform.

170

Analytics Data Integration

IN THIS SECTION:

Understand What Happens When You Enable Replication

When you enable replication, Wave looks at the sfdcDigest nodes in your existing scheduled dataflows to see which Salesforce objects and fields you’re currently extracting. Using this information, Wave enables replication for each of these objects to extract the data separately. Understanding exactly what happens when you enable replication helps you identify any additional actions that you might need to take to ensure replication runs successfully.

Enable Replication

Enable replication to decouple the extract of Salesforce data from your dataflows.

Schedule, Run, and Monitor Replication

You can schedule replication to run automatically, or run it manually, and monitor its progress, all in the data manager.

Extract Data Incrementally in Replication

By default, replication extracts data from each Salesforce object incrementally. Records are inserted, updated, or deleted to match changes in the object since the previous replication run. Because only changes are extracted, replication runs faster.

Configure Replication Settings

When a Salesforce object is enabled for replication, you can configure its replication settings to disable fields that you don’t want to replicate, add a filter to only replicate specific data, or disable replication for an object. You do all this through the data manager.

Disable Replication for an Object

Disable replication for an object in the data manager to exclude it from replication.

Add Objects and Fields to Replication

It is not possible to add objects or fields to replication in the data manager. To add an object, add an sfdcDigest node for it to one of your dataflows. To add a field for an object, add it to an existing sfdcDigest node for the object in one of your dataflows. The new object or field is added automatically to replication when you upload the updated dataflow definition file.

Considerations When Creating or Updating Dataflows with Replication Enabled

If you have enabled replication, here are some things to keep in mind the next time you create or update one of your dataflows.

Considerations When Using Wizard-Based Wave Apps with Replication

If you use wizard-based Wave Analytics apps, such as Sales or Service Wave, here are some things to consider before and after you enable replication.

Replication Limits and Limitations

Here are some things to consider when you’re working with replication.

Understand What Happens When You Enable Replication

Here’s what happens, and the action we recommend that you take.

Your Existing Dataflows Are Validated

Because replication is based on your existing dataflows, Wave validates your scheduled dataflows to ensure that they are error free.

Although dataflows with errors are not considered for replication, Wave alerts you to any errors in sfdcDigest nodes when you enable replication.

171

Analytics Data Integration

Recommended Action

Before you enable replication, review you scheduled dataflows for errors. In particular, look out for incorrect object or field names in sfdcDigest nodes. If you see errors displayed when you enable replication, fix them in your dataflows before you continue. When you fix errors in a dataflow, Wave automatically considers it for replication when you upload the dataflow definition file.

Important:

Wave does not consider unscheduled dataflows when you enable replication. Make sure that any dataflow that you

want to be considered is scheduled. See Reschedule and Unschedule a Dataflow

Your Dataflows Are Processed in Order of Dataflow Creation Date

Wave considers the sfdcDigest nodes in your dataflows in the order in which they were created. If there are multiple sfdcDigest nodes for the same object in your dataflows, they are merged. If there are conflicts with object-level parameters, the values from the most recently created dataflow are used for replication. Consider the following example dataflows, each containing an sfdcDigest node for the Opportunity object:

Dataflow A (Created 08/01/2016) Dataflow B (Created 08/15/2016)

{

"Extract_OpportunitiesA": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"incremental": false,

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "StageName" },

{ "name": "CloseDate" },

{ "name": "AccountId" },

{ "name": "OwnerId" }

]

}

{

"Extract_OpportunitiesB": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"incremental": true,

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "StageName" },

{ "name": "CloseDate" },

{ "name": "AccountId" },

{ "name": "OwnerId" }

]

}

In the resulting replication for this object, the incremental

parameter is set to true

, from the most recently created dataset. This also applies to other object-level attributes, such as fullRefreshToken

and enablePKChunking

Recommended Action

Review your dataflows to ensure that object-level parameters are consistent across all sfdcDigest nodes for the same object.

All Fields in Multiple sfdcDigest Nodes for an Object Are Included in Replication

If your dataflows have multiple sfdcDigest nodes for the same object, each with a different set of fields, a superset of all these fields is included in replication. In addition, if the same field has different attributes, such as type

or isMultivalue

, the attributes from the most recently created dataflow are used for replication. Consider the following example dataflows, each containing an sfdcDigest node for the Opportunity object:

172

Analytics

Dataflow A (Created 08/01/2016)

{

"Extract_OpportunitiesA": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id"},

{ "name": "Name" },

{ "name": "StageName" },

{ "name": "CloseDate",

"fiscalMonthOffset": 9 },

{ "name": "AccountId" },

{ "name": "OwnerId" }

]

}

Dataflow B (Created 08/15/2016)

{

"Extract_OpportunitiesA": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id"},

{ "name": "Name" },

"fiscalMonthOffset": 3 }

]

}

{ "name": "StageName" },

{ "name": "CloseDate",

}

Data Integration

In the resulting replication for this object, the Id, Name, StageName, CloseDate, AccountId, and OwnerId are included. The fiscalMonthOffset

attribute for CloseDate is set to

from the most recently created dataflow.

Recommended Action

Review the fields in the sfdcDigest nodes of your dataflows. If the resulting replication includes fields that you don’t require, you can

disable them in the replication settings for that object (see Configure Replication Settings

). If you have field attribute conflicts, ensure that the attributes you want to use for replication are in the most recently created dataflow.

Advanced Filters on sfdcDigest Nodes Are Included in Replication

If an sfdcDigest node includes a complexFilterConditions

parameter to apply an advanced filter, Wave adds this filter to the replication settings for that object. When there are multiple sfdcDigest nodes with complex filters for the same object, only the filter from the most recently created or updated dataflow is added to the replication settings. Wave ignores structured filters applied through filterConditions

parameters.

Recommended Action

After you enable replication, review each object’s replication settings to verify that filters have been applied correctly (see

Configure

Replication Settings ). To use multiple filters for an object, we recommend that you remove any filter from replication settings, and

add the filters to the corresponding dataflow using the filter transformation (see filter Transformation ).

Note:

If you have participated in the replication pilot program and have sfdcDigest node filters in your replication dataflow, these filters are not added to replication. Add these filters manually through the object’s replication settings. See

Configure Replication

Settings

SEE ALSO:

Filtering Records Extracted from a Salesforce Object

173

Analytics Data Integration

Enable Replication

Enable replication to decouple the extract of Salesforce data from your dataflows.

EDITIONS

Important: We recommend that you read

Understand What Happens When You Enable

Replication before you enable replication. This will help you identify any additional actions

that you might need to take to ensure replication runs successfully.

To enable replication, perform the following steps.

1. From Setup, enter

Wave Analytics

in the Quick Find box, then select Settings.

2. Select Enable Replication, and then click Save.

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

Schedule, Run, and Monitor Replication

Important: After you enable replication, make sure you run replication before your dataflows next run. Dataflows with sfdcDigest nodes fail until replication has run and completed for the first time. See

USER PERMISSIONS

Schedule, Run, and Monitor Replication

You can schedule replication to run automatically, or run it manually, and monitor its progress, all in the data manager.

To enable replication:

•

“Customize Application”

Important:

After you enable replication, make sure you run replication and it completes before your dataflows next run. Dataflows with sfdcDigest nodes fail until replication has run for the first time.

IN THIS SECTION:

Schedule Replication

Schedule replication to run automatically.

Run Replication

Run replication manually at any time in the data manager.

Monitor Replication

Monitor the progress of replication in the data manager.

174

Analytics Data Integration

Schedule Replication

Schedule replication to run automatically.

When you schedule replication, you schedule it to run for all your enabled objects at the same time.

You cannot schedule separate replications for each object.

Note: A replication run counts towards your limit of 24 dataflow runs per day. For example, in an organization using the Sales Wave and Service Wave apps, a user dataflow, and replication, there are four separate dataflows. If you schedule each one to run every six hours, you will reach the daily limit. Consider this limit when you schedule replication.

To schedule replication, perform the following steps.

In Wave Analytics, click the gear icon ( manager.

) and then click Data Manager to open the data

2. In the data manager, click the Replication tab.

The Replication tab displays a list of Salesforce objects enabled for replication.

3. Click Edit Replication Schedule.

The schedule dialog appears.

4. Enter the frequency and then click Save.

5. Click Done.

Run Replication

Run replication manually at any time in the data manager.

Note:

When you run replication manually, consider that it counts towards your limit of 24 dataflow runs per day.

To run replication, perform the following steps.

In Wave Analytics, click the gear icon ( manager.

) and then click Data Manager to open the data

2. In the data manager, click the Replication tab.

The Replication tab displays a list of Salesforce objects enabled for replication.

3. Click Run Replication Now.

The Monitor page of the data manager opens to display the replication of each object as a separate job.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

USER PERMISSIONS

To schedule replication:

•

“Edit Wave Analytics

Dataflows”

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

USER PERMISSIONS

To run replication:

•

“Edit Wave Analytics

Dataflows”

175

Analytics Data Integration

Monitor Replication

Monitor the progress of replication in the data manager.

To monitor replication, perform the following steps.

In Wave Analytics, click the gear icon ( manager.

) and then click Data Manager to open the data

2. In the data manager, click Monitor.

The Replication tab displays a list of Salesforce objects enabled for replication.

3. Click Run Replication Now.

The monitor page displays information from the Wave data monitor.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

Extract Data Incrementally in Replication

Note: A full extract is performed the first time replication runs for an object.

USER PERMISSIONS

To monitor replication:

•

“Edit Wave Analytics

Dataflows”

To turn off incremental mode for a replicated object, add the incremental

parameter to the object’s sfdcDigest node in your dataflow, and set its value to false

. For example:

{

"Extract_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"incremental": false,

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "StageName" },

{ "name": "CloseDate" },

{ "name": "AccountId" },

{ "name": "OwnerId" }

]

}

To force a full extract for a replicated object, add the fullRefreshToken

parameter to the object’s sfdcDigest node, and give it a random value. For example:

{

"Extract_Opportunities": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fullRefreshToken": "bb",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "StageName" },

176

Analytics Data Integration

{ "name": "CloseDate" },

{ "name": "AccountId" },

{ "name": "OwnerId" }

]

}

The next time you need to force a full extract for this object, change the value in the fullRefreshToken

parameter.

Considerations When Using Incremental Extract with Replication

•

If incremental mode is explicitly turned off in an sfdcDigest node, and not explicitly turned on in later nodes for the same object, it is not turned on when you enable replication. You have to manually turn it on by setting the incremental

parameter to true in the dataflow.

•

When you enable replication, incremental mode is not turned on for Salesforce objects that perform hard deletes on records, unless it’s explicitly turned on in the dataflow.

•

Changes made less than one hour before replication runs are not picked up. Bear this in mind if you have records that are frequently updated.

•

As a best practice, force a full refresh regularly, using the fullRefreshToken

parameter, to ensure that all changes are picked up.

•

Formula fields can become out of sync with your replicated object in incremental mode. We recommend that you turn off incremental extracts for objects containing formula fields.

•

If you make field changes in an sfdcDigest dataflow node, such as adding or removing fields, or changing attributes, Wave triggers a full extract of data the next time replication runs.

SEE ALSO:

Configure Replication Settings

To configure replication settings, perform the following steps.

In Wave Analytics, click the gear icon ( manager.

) and then click Data Manager to open the data

2. In the data manager, click the Replication tab.

The Replication tab displays a list of Salesforce objects enabled for replication.

3. Click an object to open its replication settings.

A preview of the replicated data appears.

4. To disable replication for a field, hover over the field’s column header and click the cross that appears.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

Add Objects and Fields to Replication

USER PERMISSIONS

To configure replication settings:

•

“Edit Wave Analytics

Dataflows”

177

Analytics Data Integration

Note: You can’t disable replication for a field that’s used in a dataflow, or disable replication for an object’s Id field.

To filter the data that’s replicated, click .

a. Enter a filter using a Salesforce Object Query Language (SOQL) WHERE clause expression. For example:

RecordType = 'B2B' AND Status = 'Open'

b. Click Done.

Note:

Your preview data does not reflect any filter you add. Filters are only applied when you save the replication settings.

6. Click Save to save the replication settings.

Note: If you disable replication for a field, you cannot enable it again through the replication settings. You can re-enable a field by adding it to the sfdcDigest node in your dataflow. See

Disable Replication for an Object

Disable replication for an object in the data manager to exclude it from replication.

Note:

You can’t disable replication for an object that’s used in a dataflow.

To disable replication for an object, perform the following steps.

In Wave Analytics, click the gear icon ( manager.

) and then click Data Manager to open the data

2. In the data manager, click the Replication tab.

The Replication tab displays a list of Salesforce objects enabled for replication.

Click the actions button ( ) to the right of the object and then select Disable Replication.

A warning message appears.

4. Read the warning message carefully and then click Disable.

Note:

If you disable replication for an object, you can’t enable it again in the data manager.

You can re-enable an object by adding an sfdcDigest node for it to your dataflow. See

Add

Objects and Fields to Replication .

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

USER PERMISSIONS

To disable replication for an object:

•

“Edit Wave Analytics

Dataflows”

178

Analytics Data Integration

Add Objects and Fields to Replication

SEE ALSO:

replication for an object on the Replication tab of the data manager. See Disable Replication for an Object

Considerations When Creating or Updating Dataflows with Replication Enabled

If you have enabled replication, here are some things to keep in mind the next time you create or update one of your dataflows.

•

Wave validates the dataflow definition file when you upload it. If you see errors displayed, correct them and upload the file again.

•

Dataflow definition file uploads might take longer. This is because Wave is validating the file and using the sfdcDigest nodes to define replication settings.

•

If you remove an object’s sfdcDigest node from the definition file, replication is not disabled for that object. If necessary, disable

•

If you remove a field from an sfdcDigest node in the definition file, replication is not disabled for that field. If necessary, disable

replication for the field in replication settings. See Configure Replication Settings

•

If you add a field in the definition file, that field is enabled in the replication settings.

•

If you unschedule the dataflow, Wave does not disable replication for the objects and fields.

Considerations When Using Wizard-Based Wave Apps with Replication

If you use wizard-based Wave Analytics apps, such as Sales or Service Wave, here are some things to consider before and after you enable replication.

Enabling Replication When You Are Already Using Wave Apps

If you already have Wave apps when you enable replication, it’s important that replication includes all the Salesforce objects and fields used in these apps. If it doesn’t, the affected Wave app dataflows fail. Complete these steps to ensure all the required objects and fields are included.

1. Before you enable replication, set a recurring schedule on each of your Wave app dataflows. Wave considers objects and fields in only scheduled dataflows when you first enable replication.

2. After you enable replication, check replication settings to verify that all the required objects and fields have been included. See

Configure Replication Settings .

3. Run replication to ensure that all objects and fields in the Wave app dataflows have been replicated.

4. Run your Wave app dataflows to verify that they run to completion.

5. Schedule replication to run before all your dataflows, to keep your Wave app data up to date.

Creating a New Wizard-Based Wave App After Enabling Replication

When you create a wizard-based Wave app, objects and fields from the new dataflow are automatically added to replication, if they’re not already included. Complete these steps after you create a Wave app.

1. Check replication settings to verify that all the required objects and fields have been included. See Configure Replication Settings

179

Analytics Data Integration

2. Run replication to ensure that all objects in the new dataflow have been replicated.

3. Run the new Wave app dataflow.

4. Schedule the new Wave app dataflow to run after replication.

Replication Limits and Limitations

Here are some things to consider when you’re working with replication.

•

You can enable up to 50 objects for replication. If you are currently extracting more than 50 objects in your dataflows, contact

Salesforce Customer Support before you enable replication

•

Replication only performs one extract per object. Keep this in mind if you need to perform multiple extracts from the same object.

For example, if you need to extract both filtered and unfiltered data from an object.

•

By default, replication extracts data from each object incrementally. However, changes made less than one hour before replication runs are not picked up. Keep this in mind if you have records that are frequently updated.

•

Replication counts towards your limit of 24 dataflow runs per day. For example, in an organization using the Sales Wave and Service

Wave apps, a user dataflow, and replication, there are four separate dataflows. If you schedule each one to run every six hours, you will reach the daily limit.

•

There is a known issue where replication does not generate notification emails.

Wave Connector for Excel Data

The Salesforce Wave Connector makes it easy to import data from Microsoft Excel 2013 to Wave Analytics.

IN THIS SECTION:

Install the Wave Connector Excel App

The Wave Connector app gives users a fast, easy way to import data from Excel 2013 into Salesforce Wave Analytics..

Install the Wave Connector Excel App

The Wave Connector app gives users a fast, easy way to import data from Excel 2013 into Salesforce

Wave Analytics..

USER PERMISSIONS

To import data from Excel

2013 to Wave Analytics :

•

Upload External Data to

Wave Analytics

180

Analytics Data Integration

If you use Excel 2013 on the desktop or Office 365, the Office Online version of Excel, the Wave Connector gives you a great way to get your data into Salesforce Wave Analytics. After installing the Connector, you just select data from Excel, click Submit, and the Connector does the work for you, importing the data to Wave Analytics and creating a dataset.

Here’s how to install the Connector:

1. Open Excel, either on your desktop or in Office Online.

2. Open the Insert tab.

3. Click Apps for Office.

4. Search for the Wave Connector, and click to install it.

5. Enter your Salesforce credentials to open the Connector.

Once you’ve installed the Connector, follow the instructions in the Connector window to create datasets based on Excel data. Opening the Connector automatically logs you in to Salesforce Wave Analytics. Click the Connector Help icon for complete information about using the app.

181

Analytics Data Integration

Create a Dataset with External Data

Tip: Wave temporarily stores the uploaded CSV and metadata files for processing only. After the datasets are created, Wave purges the files. If you want to use the files again later, keep a copy.

Before uploading external data files, review the format requirements and examples of the .csv and metadata files in the External Data Format Reference .

Note:

You can also use the the External Data API to upload external data files. Use the API to take advantage of additional features, like performing incremental extracts and performing append, delete, and upsert operations. For more information about the External Data API, see the External Data API Developer’s Guide .

1. On the home or app page, click Create > Dataset.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

Rules for Automatic Generation of a Metadata File

USER PERMISSIONS

To upload external data:

•

“Upload External Data to

Wave Analytics”

2. Type the name of your dataset in the Dataset Name field.

The name cannot exceed 80 characters.

3. If you want to create the dataset in a different app, change the app in the App drop-down list.

4. Click CSV.

5. Add the .csv file.

After you add the .csv file, Wave Analytics automatically generates and adds the corresponding metadata file.

182

Analytics Data Integration

Note:

Instead of using the generated metadata file, if you want, you can upload a different metadata file that you created from scratch. If you upload your own metadata file, the Preview Data button is disabled.

6. Perform the following tasks to change the metadata attributes in the generated metadata file.

a. Click Preview Data to view and change the required metadata attributes.

b. Click a column name to change it. The column name is the display name in the dataset. The column name cannot exceed 40 characters.

183

Analytics

c. Click the column header to change other attributes for the column.

You can change the attributes for measure and date columns only.

Data Integration

d. To apply the changes to all other columns of the same data type, click Apply to All <data type> Columns.

e. Click Submit to save the metadata changes to the metadata file.

Note:

If there are errors, the Submit button is grayed out.

f. Click OK to close the confirmation message.

g. To change optional metadata attributes, click

to download the metadata file, edit the file, and then upload it.

7. Click Create Dataset.

Your data files are scheduled for upload. It might take some time to process the data upload job; you can monitor its status in the data monitor. If upload is successful, the new dataset is available from the home or app page.

8. Click Continue to dismiss the confirmation message.

184

Analytics Data Integration

IN THIS SECTION:

When you upload a CSV file from the user interface, Wave Analytics automatically generates the metadata file as long as the CSV file meets certain requirements.

Monitor an External Data Upload

When you upload an external data file, Wave Analytics kicks off a job that uploads the data into the specified dataset. You can use the data monitor to monitor and troubleshoot the upload job.

Rules for Automatic Generation of a Metadata File

When you upload a CSV file from the user interface, Wave Analytics automatically generates the metadata file as long as the CSV file meets certain requirements.

To enable Wave Analytics to generate the metadata file, a CSV file must meet the following requirements.

•

The file type must be .csv, not .gz or .zip.

•

The file must contain one row for the column header and at least one record.

•

The CSV file must meet all Wave Analytics requirements as mentioned in the External Data Format Reference .

Wave Analytics generates the metadata attributes for each CSV column based on the first 100 rows in the CSV file. Wave Analytics uses the following rules to convert the CSV column names to field labels.

•

Replaces special characters and spaces with underscores. For example, "Stage Name" becomes "Stage_Name."

•

Replaces consecutive underscores with one underscore, except when column name ends with "__c." For example, "stage*&name" becomes "stage_name."

•

Prefixes the field label with "X" when the first character of the column name is numeric. For example, "30Day" becomes "X30Day."

•

Replaces the field name with "Column" + column number when all characters in the column name are not alphanumeric. For example, the fourth column name "*&^*(&*(%" becomes "Column4."

•

Deletes underscores at the beginning and end of the field label to ensure that it doesn't start or end with an underscore.

•

Increments the derived field label if the label is the same as an existing label. For example, if "X2" already exists, uses "X21," "X22,"

"X23."

Tip: You can download the generated metadata file to change the metadata settings, and then upload it to apply the changes.

You can download the metadata file when you create or edit a dataset.

185

Analytics Data Integration

Monitor an External Data Upload

When you upload an external data file, Wave Analytics kicks off a job that uploads the data into the specified dataset. You can use the data monitor to monitor and troubleshoot the upload job.

The Jobs view (1) of the data monitor shows the status, start time, and duration of each dataflow job and external data upload job. It shows jobs for the last 7 days and keeps the logs for 30 days.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

USER PERMISSIONS

To access the data monitor:

•

“Edit Wave Analytics

Dataflows,” “Upload

External Data to Wave

Analytics,” or “Manage

Wave Analytics”

In Wave Analytics, click the gear button ( ) and then click Data Monitor to open the data monitor.

The Jobs view appears by default. The Jobs view displays dataflow and upload jobs. The Jobs view displays each upload job name as

<dataset_name upload flow>

. You can hover a job to view the entire name.

Note: To view external data upload jobs in the Jobs view, Show in the File Uploads field (2) must be selected. It’s selected by default.

186

Analytics

To see the latest status of a job, click the Refresh Jobs button (

Each job can have one of the following statuses.

Status

Queued

Running

Failed

Successful

Warning

Description

The job is in queue to start.

The job is running.

The job failed.

The job completed successfully.

The job completed successfully, but some rows failed.

Data Integration

3. To view the run-time details for a job, expand the job node (3).

The run-time details display under the job. In the run-time details section, scroll to the right to view information about the rows that were processed.

4. To troubleshoot a job that has failed rows, view the error message. Also, click the download button (1) in the run-time details section to download the error log.

Note:

Only the user who uploaded the external data file can see the download button.

The error log contains a list of failed rows.

187

Analytics Data Integration

5. To troubleshoot a failed job, view the error message and the run-time details.

Edit a Dataset

USER PERMISSIONS

To view a dataset edit page:

To update a dataset name, app, and extended metadata:

To delete a dataset:

To upload and preview data:

To edit a dataset security predicate:

“Use Wave Analytics” AND Editor access to the dataset’s app

“Upload External Data to Wave Analytics”

AND Editor access to the dataset’s app

“Edit Wave Analytics Dataflows”

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

You can edit a dataset to change the dataset name, app, security predicate, or extended metadata (XMD) file associated with the dataset.

For datasets created from an external data file, you can also upload a new external data file or metadata file to update the data or metadata.

If you add an external data file, Wave Analytics generates and adds the corresponding metadata file. To make further changes to the metadata, you can click Preview Data or download and edit the generated metadata file. You can also upload your own metadata file to overwrite the generated file.

1. On the home or app page, click the Datasets tab.

2. Hover over the dataset that you want to edit, and then click Edit.

3. Configure the following options if applicable.

Option

Dataset Name

App

Add Extended Metadata File (JSON)

Description

Enter a new name if you’d like to change the name of the dataset.

The name cannot exceed 80 characters.

Select a new app if you’d like to move the dataset to a different app.

Specify an extended metadata file if you’d like to customize the formatting of dashboards associated with the dataset.

Refer to Extended Metadata (XMD) Reference for information about extended metadata files.

Add External Data File (CSV)

Specify an external data file if you’d like to replace the existing data in the dataset with data from the external data file.

Maximum file size is 500 MB. You can upload a .CSV, .GZ, or .ZIP

file.

188

Analytics

Option

Add Metadata File (JSON)

Security Predicate

Data Integration

Description

Refer to External Data Format Reference for information about external data files and metadata files.

Specify a metadata file if you’d like to redefine the structure of the external data file. If you upload a new metadata file, you must also upload the corresponding external data file.

Refer to External Data Format Reference for information about metadata files.

Add a security predicate if you’d like to apply row-level security on the dataset.

For information about predicates, see

Row-Level Security for

Datasets .

4. If you uploaded a new .csv file, click Preview Data to view and change the required metadata attributes.

You can change the optional metadata later.

Note: The Preview Data button is disabled if you uploaded your own metadata file.

After you click Preview Data, the preview page appears.

5. For each column:

a. Click a column name to change it. The column name is the display name in the dataset. The column name cannot exceed 40 characters.

189

Analytics

b. Click the column header to change other required attributes for the column.

You can change the attributes for measure and date columns only.

Data Integration

c. To apply the changes to all other columns of the same data type, click Apply to All <data type> Columns.

6. Click Submit to save the metadata changes in the preview page to the metadata file.

Note:

The Submit button is grayed out if there are errors.

7. Click OK to close the confirmation message.

8. To change optional metadata attributes—which are not visible in the preview page—click to download the metadata file, edit the file, and then upload it.

9. Click Update Dataset.

10. Click Continue to dismiss the confirmation message.

190

Analytics Data Integration

Delete a Dataset

Delete unnecessary datasets from your My Private App or in shared apps on which you have at least

Editor access. Removing datasets reduces clutter and helps you avoid reaching your org's limit for rows across registered datasets.

When you delete a dataset, Wave Analytics permanently deletes the dataset and doesn’t delete the corresponding lenses or dashboards that reference the dataset. Lenses and dashboards that reference a deleted dataset will no longer be available. As a result, Salesforce.com recommends that you remove the associated lenses and dashboards before you delete a dataset.

If a dataflow transformation —like edgemart or sfdcRegister— references the dataset, you must remove the reference before you can delete the dataset. For example, to delete the “Opportunities” dataset, you must remove the sfdcRegister transformation from the dataflow snippet shown below.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

Create a Dataset with a Recipe

USER PERMISSIONS

To delete a dataset:

•

“Use Wave Analytics”

AND Editor access to the dataset’s app

{

...

"Register_Dataset": {

"action": "sfdcRegister",

"parameters": {

"alias": "Opportunities",

"name": "Opportunities",

"source": "Extract_Opportunities"

}

...}

Warning: You can’t recover a deleted dataset.

1. On the home or app page, click the Datasets tab.

2. Hover over the dataset that you want to delete, and then click Edit.

3. Click Delete Dataset.

If applicable, Wave Analytics shows a list of all lenses and dashboards that reference the dataset and that you have access to view.

After you delete the dataset, any lens or dashboard that reference the dataset will become unusable.

4. Click Delete Permanently and confirm.

Create Datasets with Recipes in Data Prep (Beta)

Note: Data prep is currently a beta feature that is production quality but has known limitations. Contact Salesforce if you’d like to try it out.

191

Analytics Data Integration

In data prep, you create datasets using recipes. A recipe is a set of transformations, or steps, that you want to perform on a specific source dataset or replicated object. When you run a recipe, it applies the transformations and outputs the results to a new target dataset.

Here’s what you can do with your source data in a recipe:

•

Remove fields.

•

Add fields from another dataset or replicated object.

•

Transform fields.

You can run a recipe once only, or schedule it to run on a recurring basis to keep your target dataset up to date. Wave saves your recipes, so you can update them or even use them to create new recipes.

IN THIS SECTION:

Use a recipe to combine data from multiple datasets or replicated objects, choose fields, transform field values, and output the results to a new target dataset.

Start a Recipe from a Dataset or Replicated Object

Start a dataset recipe by going to the dataset or replicated object that contains the data you want to start from.

Set Preview Data in a Dataset Recipe

A dataset recipe displays a preview of your data as you work. You can set the fields and number of rows that appear in this preview.

If you are working with a large set of data, reducing the size of the preview speeds up the data refresh as you add steps to your recipe.

Add Data in a Dataset Recipe

Combine data in a recipe by adding fields from other datasets or replicated objects.

Transform Fields in a Dataset Recipe

Data is not always consistent, especially when you combine data from different sources. In a dataset recipe, you can transform fields to ensure that values are consistent in your target dataset. Change case, split or truncate values to get just the parts you need, and replace incorrect values.

Navigate Fields in a Dataset Recipe

As you add data and transform fields, the number of fields in your recipe preview increases. This makes it harder to find the fields that you want to work with. Use the field navigator to quickly find the fields that you need, and hide the ones you don't. The navigator also shows you the attributes of a field, such as its API name, and field type.

Save a Dataset Recipe

If you’re not ready to run a recipe, save it and come back to it later. Saving simply saves your steps without you having to run the recipe.

192

Analytics Data Integration

Run a Dataset Recipe

Run a new recipe to create a target dataset, or run an existing recipe to update a target dataset.

Manage and Monitor Dataset Recipes

View the status of your recipes and target databases on the Recipes tab of the data manager. You can also open recipes to updated or run them, or delete them.

Create a Dataset with a Recipe

Use a recipe to combine data from multiple datasets or replicated objects, choose fields, transform field values, and output the results to a new target dataset.

Tip:

Before you get started, it’s a good idea to think about the requirements for the target dataset you want to create. Consider these questions:

•

Does the data you want to work with exist in a dataset or as a replicated object?

•

If you are combining data, how is the data related?

•

What fields do you want in the target dataset?

•

Do you need to transform any field values?

•

In what app are you going to create the target dataset?

•

Do you need to control access to the data in the target dataset?

When you’re ready, follow these steps to create a dataset with a recipe.

Start a Recipe from a Dataset or Replicated Object

Start a dataset recipe by going to the dataset or replicated object that contains the data you want to start from.

IN THIS SECTION:

Start a Recipe from a Dataset

To start a recipe from a dataset, perform the following steps.

Start a Recipe from a Replicated Object

To start a recipe from a replicated object, perform these steps.

193

Analytics Data Integration

Start a Recipe from a Dataset

To start a recipe from a dataset, perform the following steps.

In Wave Analytics, click the gear icon ( manager.

) and then click Data Manager to open the data

2. In the data manager, click the Datasets tab.

The Datasets tab displays a list of your available datasets.

Either click the object you require, or click the actions button ( ) to the right of the object and then select Create New Recipe.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

USER PERMISSIONS

To create a recipe:

•

“Edit Wave Analytics

Dataflows”

You are prompted to enter a name for the recipe.

4. Enter a recipe name and then click Continue.

Note:

The recipe name you enter here is also used for the target dataset name.

The dataset recipe page opens, displaying a preview of the data from your dataset.

194

Analytics Data Integration

Start a Recipe from a Replicated Object

To start a recipe from a replicated object, perform these steps.

In Wave Analytics, click the gear icon ( manager.

) and then click Data Manager to open the data

2. In the data manager, click the Replication tab.

The Replication tab displays a list of Salesforce objects enabled for replication.

Click the actions button ( ) to the right of the required object and then select Create New

Recipe.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

USER PERMISSIONS

To create a recipe:

•

“Edit Wave Analytics

Dataflows”

You are prompted to enter a name for the recipe.

4. Enter a recipe name and then click Continue.

Note: The recipe name you enter here is also used for the target dataset name.

The dataset recipe page opens, displaying a preview of the data from your dataset.

Set Preview Data in a Dataset Recipe

A dataset recipe displays a preview of your data as you work. You can set the fields and number of rows that appear in this preview. If you are working with a large set of data, reducing the size of the preview speeds up the data refresh as you add steps to your recipe.

As you work in a dataset recipe, the size of your preview is always visible at the top of the page.

195

Analytics

Click this preview information to view or change what you see in the preview.

Data Integration

Here’s what you can do to set your recipe preview data:

1. Can’t see a field in the list? Start typing the field name here to narrow down the list.

2. Not sure if you need a field in your preview? Click a field in the list to see a preview of its values here.

3. Don’t need field in your preview? Click the checkbox to remove it.

4. Want fewer rows in your preview? Choose to see the first 1000, 5000, or 10000 rows and select a field to sort them by.

Click Apply when you’re done, to return to the recipe.

Note: If you remove a field from the recipe preview, by default it is not included when the recipe runs. However, you can reselect the field when you run the recipe.

196

Analytics Data Integration

Add Data in a Dataset Recipe

Combine data in a recipe by adding fields from other datasets or replicated objects.

When you add data to a recipe, you select the lookup dataset or replicated object that contains the fields you want to add. You also select the merge key fields to match the records in the existing and new data. For example, to add User fields to Opportunity records, use the User Id and

Opportunity Owner Id fields as your merge keys.

Let’s look at another example here, to illustrate how you add data. Imagine you have started a recipe from a US Leads dataset, and you want to add demographic data based on zip code from another dataset. You can use the Zip Code field from each dataset as the merge keys.

On the dataset recipe page, click the Add Data button ( ).

2. Select either the Datasets or Replication tab, and click the dataset or object you want to add.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

Manage and Monitor Dataset Recipes .

USER PERMISSIONS

To create a recipe:

•

“Edit Wave Analytics

Dataflows”

3. Click a column header to select it as the merge key field in the origin dataset, and do the same in the lookup dataset.

Then click Continue.

Important: Wave datasets contain date and time component fields, such as Year, Quarter, and Month, that are derived automatically from existing date and time fields. Use of these component fields as merge keys is not supported.

4. Click the column header or column checkbox to select each field that you want to add.

197

Analytics Data Integration

Then click Continue.

Each time you add data, a new step is created in your recipe, and the added fields appear in your recipe preview. Click the Navigate

Recipe button (1) to view your recipe steps. To undo a step, click the actions button (2) and select Remove Transformation.

Transform Fields in a Dataset Recipe

When you transform a field, a new field is added with the new values to the right of the original field. You can choose to keep or remove the original field.

To transform a field, click the field header (1), and then click the Transform Suggestions button (2). Your transform options appear in the side panel (3). Click Apply to apply a transform. When you’re finished, click another field header to move on to another field.

198

Analytics Data Integration

These are the field transforms you can apply.

Transform Description

Convert values in selected field to uppercase.

Example

Correct US state values.

becomes

Convert values in selected field to lowercase.

Correct lead email addresses captured at a trade show.

[email protected]

becomes

[email protected]

Divide values into multiple parts at the delimiter and create a new field for each part.

Divide phone numbers into area code and number.

Select the

delimiter to split

925-900 2123

into two new values

925

and

900 2123

Starting at the character Position in a value, extract the number of characters in Length into a new field.

Extract the month from date values.

Enter a Position of

and a Length of

to extract the month value

from the month value

2016-10-23

Find a specific value in a field and replace it with a new value. A new field is created containing both replaced and unchanged values.

Make country values consistent.

Find all instances of

and Replace them with

USA

Important:

Wave datasets contain date and time component fields, such as Year, Quarter, and Month, that are derived automatically from existing date and time fields. Transformations on these component fields are not supported.

199

Analytics Data Integration

Navigate Fields in a Dataset Recipe

To open the field navigator in a recipe, click the Navigate Fields button ( ).

To quickly find a field, start typing its name in the Search Fields box (1) to narrow down the list of fields. When you see the field that you want, click it (2) to see it highlighted in the recipe preview. To learn more about the field’s metadata, click Attributes (3). If you decide you don’t need to see the field, click the eye (4) to hide it. Click the eye again to show a field. You can also hide a field by hovering over its header in the preview, and clicking the eye.

Note: If you hide a field in the navigator, by default it is not included when the recipe runs. However, you can choose to include the field when you run the recipe.

Save a Dataset Recipe

If you’re not ready to run a recipe, save it and come back to it later. Saving simply saves your steps without you having to run the recipe.

To save a recipe, click Save.

200

Analytics Data Integration

You have two options when saving a recipe.

Save Recipe

Save the recipe with the name you provided when you started. The recipe does not run. Use this option to save your progress.

Save As New...

Save a copy of your recipe with the same name you provided when you started. The copy opens in a new tab in the data manager.

Use this option to try out different transformations in a recipe, without changing your original recipe.

Important: A recipe you create with the Save As New option has the same recipe name and target dataset name as the original recipe. We recommend that you run either the original recipe or the new recipe, but not both, to avoid creating target datasets with the same name.

Open a saved recipe on the Recipes tab of the data manager to update it or run it. See

Run a Dataset Recipe

Run a new recipe to create a target dataset, or run an existing recipe to update a target dataset.

When you run a recipe, you select an app for the target dataset, enter a security predicate if necessary, and select the fields to include in your target dataset. You can run a recipe once only, or set it to run on a schedule.

1. On the dataset recipe page, click Create Dataset.

The Run Recipe dialog appears.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

Row-Level Security for Datasets .

USER PERMISSIONS

To create a recipe:

•

“Edit Wave Analytics

Dataflows”

2. From the App picklist, select where you want to create the target dataset.

3. If required, enter a security predicate for the target dataset. See

4. In the list of fields, make sure the fields that you require are selected.

Note: By default, fields that you have removed or hidden in the recipe preview are not selected. Make sure that you review these fields.

5. Click Continue.

201

Analytics

The schedule dialog appears.

Data Integration

6. Either select to run on a schedule and enter a frequency, or select to run once only.

7. Click Run Dataset Recipe.

Your recipe is queued to run.

Manage and Monitor Dataset Recipes

View the status of your recipes and target databases on the Recipes tab of the data manager. You can also open recipes to updated or run them, or delete them.

To open the data manager in Wave Analytics, click the gear icon ( ) and then click Data Manager.

The Recipes tab in the data manager displays your recipes and their associated target datasets.

A recipe’s icon and status indicate its current state.

Icon Status

Pushed

Saved

202

Explanation

New or updated recipe that has been run.

New or updated recipe that has been saved but not run.

Analytics

A target dataset’s icon and status indicate its current state.

Icon Status

Up to Date

Not Up to Date

Not Created

Data Integration

Explanation

The dataset has been created and is up to date.

The dataset has been created, but is not currently up to date. This is because its recipe has been updated but not run, or is queued to run.

The dataset has not been created. This is because its recipe has been created but never run, or is queued to be run.

Open a recipe to update it or run it. Either click the recipe or open the actions menu and select Open Recipe.

You can also delete a recipe from here.

Note: Deleting a recipe does not delete its target dataset.

Click Monitor to view the progress of recipes that you have run. Recipes show as separate jobs in the Jobs View of the monitor.

Row-Level Security for Datasets

To implement row-level security, define a predicate for each dataset on which you want to restrict access to records. A predicate is a filter condition that defines row-level access to records in a dataset.

203

Analytics Data Integration

When a user submits a query against a dataset that has a predicate, Wave Analytics checks the predicate to determine which records the user has access to. If the user doesn’t have access to a record, Wave Analytics does not return that record.

The predicate is flexible and can model different types of security policies. For example, you can create predicates based on:

•

Record ownership. Enables each user to view only records that they own.

•

Management visibility. Enables each user to view records owned or shared by their subordinates based on a role hierarchy.

•

Team or account collaboration. Enables all members of a team, like an opportunity team, to view records shared with the team.

•

Combination of different security requirements. For example, you might need to define a predicate based on the Salesforce role hierarchy, teams, and record ownership.

The type of security policy you implement depends on how you want to restrict access to records in the dataset.

Warning: If row-level security isn’t applied to a dataset, any user that has access to the dataset can view all records in the dataset.

IN THIS SECTION:

Considerations when Defining a Predicate for a Dataset

Applying a predicate to a dataset is more than just defining the predicate expression. You also need to consider how the predicate is dependent on the information in the dataset and where to define the predicate expression.

Row-Level Security Example based on Record Ownership

Let’s look at an example where you create a dataset based on a CSV file and then implement row-level security based on record ownership. In this example, you will create a dataset that contains sales targets for account owners. To restrict access on each record in the dataset, you will create a security policy where each user can view only sales targets for accounts that they own. This process requires multiple steps that are described in the sections that follow.

Row-Level Security Example based on Opportunity Teams

To restrict access on each record in the dataset, you will create a security policy where only opportunity members can view their opportunity. This process requires multiple steps that are described in the sections that follow.

Row-Level Security Example based on Role Hierarchy and Record Ownership

Let’s look at an example where you create a dataset based on Salesforce data and then implement row-level security based on the

Salesforce role hierarchy and record ownership. In this example, you will create a dataset that contains all opportunities. To restrict access on each record in the dataset, you will create a security policy where each user can view only opportunities that they own or that are owned by their subordinates based on the Salesforce role hierarchy. This process requires multiple steps that are described in the sections that follow.

SEE ALSO:

sfdcRegister Transformation

sfdcRegister Parameters

Considerations when Defining a Predicate for a Dataset

You can create a predicate expression based on information in the dataset. For example, to enable each user to view only dataset records that they own, you can create a predicate based on a dataset column that contains the owner for each record. If needed, you can load additional data into a dataset required by the predicate.

204

Analytics Data Integration

The location where you define the predicate varies.

•

To apply a predicate on a dataset created from a dataflow, add the predicate in the rowLevelSecurityFilter field of the Register transformation. The next time the dataflow runs, Wave Analytics will apply the predicate.

•

To apply a predicate on a dataset created from an external data file, define the predicate in the rowLevelSecurityFilter field in the metadata file associated with the external data file. Wave Analytics applies the predicate when you upload the metadata file and external data file. If you already created the dataset from a external data file, you can edit the dataset to apply or change the predicate.

Row-Level Security Example based on Record Ownership

Let’s look at an example where you create a dataset based on a CSV file and then implement row-level security based on record ownership.

In this example, you will create a dataset that contains sales targets for account owners. To restrict access on each record in the dataset, you will create a security policy where each user can view only sales targets for accounts that they own. This process requires multiple steps that are described in the sections that follow.

Note: Although this example is about applying a predicate to a dataset created from a CSV file, this procedure can also be applied to a dataset that is created from Salesforce data.

IN THIS SECTION:

Determine Which Data to Include in the Dataset

First, determine what data you want to include in the dataset. For this example, you will create a Targets dataset that contains all sales targets.

Determine Row-Level Security for Dataset

Now it’s time to think about row-level security. How will you restrict access to each record in this dataset?

Add the Predicate to the Metadata File

For a dataset created from a CSV file, you can specify the predicate in the metadata file associated with the CSV file or when you edit the dataset.

Create the Dataset

Now that you updated the metadata file with the predicate, you can create the dataset.

Test Row-Level Security for the Dataset

You must verify that the predicate is applied properly and that each user can see their own sales targets.

Determine Which Data to Include in the Dataset

First, determine what data you want to include in the dataset. For this example, you will create a Targets dataset that contains all sales targets.

You will obtain sales targets from the CSV file shown below.

AccountOwner

Tony Santos

Lucy Timmer

Bill Rolley

Keith Laz

Region

Midwest

Northeast

Midwest

Southwest

Target

10000

50000

15000

35000

TargetDate

1/1/2011

12/1/2013

1/1/2011

205

Analytics Data Integration

AccountOwner

Lucy Timmer

Region

Southeast

Target

40000

TargetDate

1/1/2011

If you were to create the dataset without implementing row-level security, any user that had access to the dataset would be able to see the sales targets for all account owners. For example, as shown below, Keith would be able to view the sales targets for all account owners.

You need to apply row-level security to restrict access to records in this dataset.

Determine Row-Level Security for Dataset

Now it’s time to think about row-level security. How will you restrict access to each record in this dataset?

You decide to implement the following predicate on the dataset.

'AccountOwner' == "$User.Name"

Note: All predicate examples in this document escape the double quotes because it’s required when you enter the predicate in the Register transformation or metadata file.This predicate implements row-level security based on record ownership. Based on the predicate, Wave Analytics returns a sales target record when the user who submits the query on the dataset is the account owner.

Let’s take a deeper look into the predicate expression:

•

AccountOwner refers to the dataset column that stores the full name of the account owner for each sales target.

•

$User.Name refers to the Name column of the User object that stores the full name of each user. Wave Analytics performs a lookup to get the full name of the user who submits each query.

Note:

The lookup returns a match when the names in AccountOwner and $User.Name match exactly—they must have the same case.

206

Analytics Data Integration

Add the Predicate to the Metadata File

For a dataset created from a CSV file, you can specify the predicate in the metadata file associated with the CSV file or when you edit the dataset.

You must escape the double quotes around string values when entering a predicate in the metadata file.

In this example, you add the predicate to the metadata file shown below.

{

"fileFormat": {

"charsetName": "UTF-8",

"fieldsDelimitedBy": ",",

"fieldsEnclosedBy": "\"",

"numberOfLinesToIgnore": 1 },

"objects": [

{

"name": "Targets",

"fullyQualifiedName": "Targets",

"label": "Targets",

"rowLevelSecurityFilter": "'AccountOwner' == \"$User.Name\"",

"fields": [

{

"name": "AccountOwner",

"fullyQualifiedName": "Targets.AccountOwner",

"label": "Account Owner",

"type": "Text"

{

"name": "Region",

"fullyQualifiedName": "Targets.Region",

"label": "Region",

"type": "Text"

{

"name": "Target",

"fullyQualifiedName": "Targets.Target",

"label": "Target",

"type": "Numeric",

"precision": 16,

"scale": 0,

"defaultValue": "0",

"format": null

{

"name": "TargetDate",

"fullyQualifiedName": "Targets.TargetDate",

"label": "TargetDate",

"description": "",

"type": "Date",

"format": "dd/MM/yy HH:mm:ss",

"isSystemField": false,

"fiscalMonthOffset": 0

}

]

}

207

Analytics Data Integration

}

]

Create the Dataset

Now that you updated the metadata file with the predicate, you can create the dataset.

Warning: If you wish to perform the steps in this sample implementation, perform the steps in a non-production environment. Ensure that these changes do not impact other datasets that you already created.

To create the dataset, perform the following steps.

1. In Wave Analytics, go to the home page.

2. Click Create > Dataset

3. Click CSV.

The following screen appears.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

USER PERMISSIONS

To upload a CSV and metadata file:

•

“Upload External Data to

Wave Analytics”

4. Select the CSV file and metadata (schema) file.

5. In the Dataset Name field, enter “SalesTarget” as the name of the dataset.

6. Optionally, choose a different app where you want to store the dataset.

7. Click Create Dataset.

208

Analytics Data Integration

Wave Analytics confirms that the upload is successful and then creates a job to create the dataset. You can view the SalesTarget dataset after the job completes successfully.

8. To verify that the job completes successfully, perform the following steps:

Click the gear icon ( ) and then select Data Monitor to open the data monitor.

By default, the Jobs View of the data monitor appears. It shows the statuses of dataflow and external data upload jobs.

Click the Refresh Jobs button ( ) to view the latest statuses of the jobs.

Test Row-Level Security for the Dataset

You must verify that the predicate is applied properly and that each user can see their own sales targets.

1. Log in to Wave Analytics as Keith.

2. Open the SalesTargets dataset.

As shown in the following lens, notice that Keith can see only his sales target.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

Determine Which Data to Include in the Dataset

Row-Level Security Example based on Opportunity Teams

Let’s look at an example where you create a dataset based on Salesforce data and then implement row-level security based on an opportunity team. In this example, you will create a dataset that contains only opportunities associated with an opportunity team. To restrict access on each record in the dataset, you will create a security policy where only opportunity members can view their opportunity.

This process requires multiple steps that are described in the sections that follow.

IN THIS SECTION:

First, determine what data you want to include in the dataset. For this example, you will create an OppTeamMember dataset that contains only opportunities associated with an opportunity team.

209

Analytics Data Integration

Design the Dataflow to Load the Data

Now it’s time to figure out how the dataflow will extract the Salesforce data and load it into a dataset. You start by creating this high-level design for the dataflow.

Determine Row-Level Security for the Dataset

Now it’s time to think about row-level security. How will you restrict access to each record in this dataset?

Modify the Dataflow Based on Row-Level Security

It’s now time to add the predicate in the dataflow definition file.

Create the Dataset

Now that you have the final dataflow definition file, you can create the dataset.

Test Row-Level Security for the Dataset

You must verify that the predicate is applied properly and that each user can see the appropriate opportunities.

Determine Which Data to Include in the Dataset

First, determine what data you want to include in the dataset. For this example, you will create an

OppTeamMember dataset that contains only opportunities associated with an opportunity team.

You will obtain opportunities from the Opportunity object and the opportunity teams from the

OpportunityTeamMember object. Both are Salesforce objects.

In this example, your Salesforce organization has the following opportunity team and users.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

Your organization also contains the following opportunities, most of which are owned by Keith.

210

Analytics Data Integration

Acc - 1000 Widgets is the only opportunity shared by an opportunity team. Bill is the Sales Manager for this opportunity. Tony is the opportunity owner.

Design the Dataflow to Load the Data

Now it’s time to figure out how the dataflow will extract the Salesforce data and load it into a dataset.

You start by creating this high-level design for the dataflow.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

The dataflow will extract data from the Opportunity and OpportunityTeamMember objects, join the data, and then load it into the OppTeamMember dataset.

Now let’s implement that design in JSON, which is the format of the dataflow definition file. A dataflow definition file contains transformations that extract, transform, and load data into a dataset.

Based on the design, you create the JSON shown below.

{

"Extract_OpportunityTeamMember": {

"action": "sfdcDigest",

"parameters": {

"object": "OpportunityTeamMember",

"fields": [

{ "name": "Name" },

{ "name": "OpportunityId" },

{ "name": "UserId" }

]

}

"Extract_Opportunity": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "AccountId" },

{ "name": "OwnerId" }

]

}

"Augment_OpportunityTeamMember_Opportunity": {

"action": "augment",

"parameters": {

"left": "Extract_OpportunityTeamMember",

211

Analytics Data Integration

"left_key": [

"OpportunityId"

"relationship": "TeamMember",

"right": "Extract_Opportunity",

"right_key": [

"Id"

"right_select": [

"Name","Amount"

]

}

"Register_Dataset": {

"action": "sfdcRegister",

"parameters": {

"alias": "OppTeamMember",

"name": "OppTeamMember",

"source": "Augment_OpportunityTeamMember_Opportunity",

"rowLevelSecurityFilter": ""

}

If you were to run this dataflow, Wave Analytics would generate a dataset with no row-level security. As a result, any user that has access to the dataset would be able to see the opportunity shared by the opportunity team.

For example, as shown below, Lucy would be able to view the opportunity that belongs to an opportunity team of which she is not a member.

You need to apply row-level security to restrict access to records in this dataset.

212

Analytics Data Integration

Determine Row-Level Security for the Dataset

Now it’s time to think about row-level security. How will you restrict access to each record in this dataset?

You decide to implement the following predicate on the dataset.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

'UserId' == "$User.Id"

This predicate compares the UserId column in the dataset against the ID of the user running a query against the dataset. The UserId column in the dataset contains the user ID of the team member associated with each opportunity. To determine the ID of the user running the query, Wave Analytics looks up the ID of the user making the query in the User object.

For each match, Wave Analytics returns the record to the user.

Modify the Dataflow Based on Row-Level Security

It’s now time to add the predicate in the dataflow definition file.

You add the predicate to the Register transformation that registers the OppTeamMember dataset as shown below.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

{

"Extract_OpportunityTeamMember": {

"action": "sfdcDigest",

"parameters": {

"object": "OpportunityTeamMember",

"fields": [

{ "name": "Name" },

{ "name": "OpportunityId" },

{ "name": "UserId" }

]

}

"Extract_Opportunity": {

"action": "sfdcDigest",

"parameters": {

213

Analytics

}

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "AccountId" },

{ "name": "OwnerId" }

]

}

"Augment_OpportunityTeamMember_Opportunity": {

"action": "augment",

"parameters": {

"left": "Extract_OpportunityTeamMember",

"left_key": [

"OpportunityId"

"relationship": "TeamMember",

"right": "Extract_Opportunity",

"right_key": [

"Id"

"right_select": [

"Name","Amount"

]

}

"Register_Dataset": {

"action": "sfdcRegister",

"parameters": {

"alias": "OppTeamMember",

"name": "OppTeamMember",

"source": "105_Augment_OpportunityTeamMember_Opportunity",

"rowLevelSecurityFilter": "'UserId' == \"$User.Id\""

}

Data Integration

214

Analytics Data Integration

Create the Dataset

Now that you have the final dataflow definition file, you can create the dataset.

Warning:

If you wish to perform the steps in this sample implementation, verify that you have all required Salesforce objects and fields, and perform the steps in a non-production environment. Ensure that these changes do not impact other datasets that you already created.

Also, always make a backup of the existing dataflow definition file before you make changes because you cannot retrieve old versions of the file.

To create the dataset, perform the following steps.

In Wave Analytics, click the gear icon ( ) and then select Data Monitor to open the data monitor.

The Jobs view of the data monitor appears by default.

2. Select Dataflow View.

3. Click the actions list (1) for the dataflow and then select Download to download the existing dataflow definition file.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

5. Add the JSON determined in the previous step

USER PERMISSIONS

To download, upload, run, and monitor a dataflow:

•

“Edit Wave Analytics

Dataflows”

4. Open the dataflow definition file in a JSON or text editor.

6. Before you save the dataflow definition file, use a JSON validation tool to verify that the JSON is valid.

An error occurs if you try to upload the dataflow definition file with invalid JSON. You can find JSON validation tool on the internet.

7. Save and close the dataflow definition file.

8. In the Dataflow View of the data monitor, click the actions list for the dataflow and then select Upload.

9. Select the updated dataflow definition file and click Upload.

10. In the Dataflow View of the data monitor, click the actions list for the dataflow and then select Run to run the dataflow job.

11.

Click the Refresh Jobs button ( ) to view the latest status of the dataflow job.

You can view the OppTeamMember dataset after the dataflow job completes successfully.

Note: If you are adding a predicate to a dataset that was previously created, each user must log out and log back in for the predicate to take effect.

215

Analytics Data Integration

Test Row-Level Security for the Dataset

You must verify that the predicate is applied properly and that each user can see the appropriate opportunities.

1. Log in to Wave Analytics as Lucy.

2. Open the OppTeamMember opportunity.

Notice that Lucy can’t view the opportunity associated with the opportunity team anymore because she is not a member of the team.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

Determine Which Data to Include in the Dataset

3. Log out and now log in as Bill.

Bill can view the opportunity that is shared by the opportunity team of which he is a member.

216

Analytics Data Integration

Row-Level Security Example based on Role Hierarchy and Record Ownership

Let’s look at an example where you create a dataset based on Salesforce data and then implement row-level security based on the

IN THIS SECTION:

First, determine what data you want to include in the dataset. For this example, you will create the OppRoles dataset that contains all opportunities as well as user details about each opportunity owner, such as their full name, division, and title.

Design the Dataflow to Load the Data

Now it’s time to figure out how the dataflow will extract the data and load it into a dataset. You start by creating this high-level design for the dataflow.

Determine Row-Level Security for the Dataset

Now it’s time to think about row-level security. How will you restrict access to each record in this dataset?

Modify the Dataflow Based on Row-Level Security

Now it’s time to modify the dataflow definition file to account for the predicate.

Create the Dataset

Now that you have the final dataflow definition file, you can create the dataset.

Test Row-Level Security for the Dataset

You must verify that the predicate is applied properly and that each user can see the appropriate opportunities.

SEE ALSO:

flatten Parameters

Determine Which Data to Include in the Dataset

First, determine what data you want to include in the dataset. For this example, you will create the

OppRoles dataset that contains all opportunities as well as user details about each opportunity owner, such as their full name, division, and title.

You will obtain opportunities from the Opportunity object and user details from the User object.

Both are objects in Salesforce.

In this example, your Salesforce organization has the following role hierarchy and users.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

217

Analytics

Also, your organization contains the following opportunities, most of which are owned by Keith.

Data Integration

Design the Dataflow to Load the Data

Now it’s time to figure out how the dataflow will extract the data and load it into a dataset. You start by creating this high-level design for the dataflow.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

The dataflow will extract data from the Opportunity and User objects, join the data, and then load it into the OppRoles dataset.

Based on the design, you create the JSON shown below.

{

"Extract_Opportunity": {

"action": "sfdcDigest",

218

Analytics

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "AccountId" },

]

{ "name": "OwnerId" }

}

"Extract_User": {

"action": "sfdcDigest",

"parameters": {

"object": "User",

"fields": [

{ "name": "Id" },

{ "name": "Username" },

{ "name": "LastName" },

{ "name": "FirstName" },

{ "name": "Name" },

{ "name": "CompanyName" },

{ "name": "Division" },

{ "name": "Department" },

{ "name": "Title" },

{ "name": "Alias" },

{ "name": "CommunityNickname" },

{ "name": "UserType" },

{ "name": "UserRoleId" }

]

}

"Augment_Opportunity_User": {

"action": "augment",

"parameters": {

"left": "Extract_Opportunity",

"left_key": [

"OwnerId"

"right": "Extract_User",

"relationship": "Owner",

"right_select": [

"Name"

"right_key": [

"Id"

]

}

"Register": {

"action": "sfdcRegister",

"parameters": {

"alias": "OppRoles",

219

Data Integration

Analytics Data Integration

"name": "OppRoles",

"source": "Augment_Opportunity_User",

"rowLevelSecurityFilter": ""

}

If you were to run this dataflow, Wave Analytics would generate a dataset with no row-level security. As a result, any user that has access to the dataset would be able to view all opportunities. For example, as shown below, Bill would be able to view all opportunities, including those owned by his manager Keith.

You need to apply row-level security to restrict access to records in this dataset.

Determine Row-Level Security for the Dataset

Now it’s time to think about row-level security. How will you restrict access to each record in this dataset?

You decide to implement the following predicate on the dataset.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

'ParentRoleIDs' == "$User.UserRoleId" || 'OwnerId' == "$User.Id"

220

Analytics Data Integration

Note: The current dataflow doesn’t contain logic to create a dataset column named “ParentRoleIDs.” ParentRoleIDs is a placeholder for the name of a column that will contain this information. In the

next step

, you will modify the dataflow to add this column to the dataset. This column name will change based on how you configure the dataflow.

Based on the predicate, Wave Analytics returns an opportunity record if:

•

The user who submits the query is a parent of the opportunity owner based on the Salesforce role hierarchy. Wave Analytics determines this based on their role IDs and the role hierarchy.

•

Or, the user who submits the query on the dataset is the opportunity owner.

Let’s examine both parts of this predicate.

Predicate Part

'ParentRoleIDs' == "$User.UserRoleId"

Description

•

ParentRoleIDs refers to a dataset column that contains a comma-separated list of role IDs of all users above the opportunity owner based on the role hierarchy. You will create this dataset column in the next section.

•

$User.UserRoleId refers to the UserRoleId column of the User object. Wave Analytics looks up the user role ID of the user who submits the query from the User object.

'OwnerId' == "$User.Id"

•

OwnerId refers to the dataset column that contains the user

ID of the owner of each opportunity.

•

$User.Id refers to the Id column of the User object. Wave

Analytics looks up the user ID of the user who submits the query from the User object.

Modify the Dataflow Based on Row-Level Security

Now it’s time to modify the dataflow definition file to account for the predicate.

In this scenario, you have to make changes to the dataflow based on the predicate.

•

Add a column in the dataset that stores a comma-separated list of the role IDs of all parents for each opportunity owner. When you defined the predicate in the previous step, you temporarily referred to this column as “ParentRoleIDs.” To add the column, you redesign the dataflow as shown in the following diagram:

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

221

Analytics Data Integration

The new dataflow design contains the following changes:

–

Extracts the role IDs from the UserRole object.

–

Uses the Flatten transformation to generate a column that stores a comma-separated list of the role IDs of all parents of each user. When you determined the predicate in the previous step, you temporarily referred to this column as “ParentRoleIDs.”

–

Link the new column to the OppRoles dataset.

•

Add the predicate to the Register transformation that registers the OppRoles dataset.

You modify the dataflow as shown below.

{

"Extract_Opportunity": {

"action": "sfdcDigest",

"parameters": {

"object": "Opportunity",

"fields": [

{ "name": "Id" },

{ "name": "Name" },

{ "name": "Amount" },

{ "name": "StageName" },

{ "name": "AccountId" },

{ "name": "OwnerId" }

]

}

"Extract_User": {

"action": "sfdcDigest",

"parameters": {

"object": "User",

"fields": [

{ "name": "Id" },

{ "name": "Username" },

{ "name": "LastName" },

{ "name": "FirstName" },

{ "name": "Name" },

{ "name": "CompanyName" },

{ "name": "Division" },

{ "name": "Department" },

{ "name": "Title" },

{ "name": "Alias" },

{ "name": "CommunityNickname" },

222

Analytics

{ "name": "UserType" },

{ "name": "UserRoleId" }

]

}

"Extract_UserRole": {

"action": "sfdcDigest",

"parameters": {

"object": "UserRole",

"fields": [

{ "name": "Id" },

{ "name": "ParentRoleId" },

{ "name": "RollupDescription" },

{ "name": "OpportunityAccessForAccountOwner" },

{ "name": "CaseAccessForAccountOwner" },

{ "name": "ContactAccessForAccountOwner" },

{ "name": "ForecastUserId" },

{ "name": "MayForecastManagerShare" },

{ "name": "LastModifiedDate" },

{ "name": "LastModifiedById" },

{ "name": "SystemModstamp" },

{ "name": "DeveloperName" },

{ "name": "PortalAccountId" },

{ "name": "PortalType" },

{ "name": "PortalAccountOwnerId" }

]

}

"Flatten_UserRole": {

"action": "flatten",

"parameters": {

"multi_field": "Roles",

"parent_field": "ParentRoleId",

"path_field": "RolePath",

"self_field": "Id",

"source": "Extract_UserRole"

}

"Augment_User_FlattenUserRole": {

"action": "augment",

"parameters": {

"left": "Extract_User",

"left_key": [

"UserRoleId"

"relationship": "Role",

"right": "Flatten_UserRole",

"right_key": [

"Id"

"right_select": [

"Roles",

"RolePath"

]

223

Data Integration

Analytics Data Integration

}

"Augment_Opportunity_UserWithRoles": {

"action": "augment",

"parameters": {

"left": "Extract_Opportunity",

"left_key": [

"OwnerId"

"right": "Augment_User_FlattenUserRole",

"relationship": "Owner",

"right_select": [

"Name",

"Role.Roles",

"Role.RolePath"

"right_key": [

"Id"

]

}

"Register": {

}

"action": "sfdcRegister",

"parameters": {

"alias": "OppRoles",

"name": "OppRoles",

== \"$User.Id\""

}

"source": "Augment_Opportunity_UserWithRoles",

"rowLevelSecurityFilter": "'Owner.Role.Roles' == \"$User.UserRoleId\" || 'OwnerId'

Note:

In this example, the dataset has columns Owner.Role.Roles and OwnerId. A user can view the values of these columns for each record to which they have access.

224

Analytics Data Integration

Create the Dataset

Now that you have the final dataflow definition file, you can create the dataset.

Warning:

Also, always make a backup of the existing dataflow definition file before you make changes because you cannot retrieve old versions of the file.

To create the dataset, perform the following steps.

In Wave Analytics, click the gear icon ( ) and then select Data Monitor to open the data monitor.

The Jobs View of the data monitor appears by default.

2. Select Dataflow View.

3. Click the actions list (1) for the dataflow and then select Download to download the existing dataflow definition file.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer

5. Add the JSON determined in the previous step

USER PERMISSIONS

To download, upload, run, and monitor a dataflow:

•

“Edit Wave Analytics

Dataflows”

4. Open the dataflow definition file in a JSON or text editor.

6. Before you save the dataflow definition file, use a JSON validation tool to verify that the JSON is valid.

An error occurs if you try to upload the dataflow definition file with invalid JSON. You can find JSON validation tool on the internet.

7. Save and close the dataflow definition file.

8. In the Dataflow View of the data monitor, click the actions list for the dataflow and then select Upload.

9. Select the updated dataflow definition file and click Upload.

10. In the Dataflow View of the data monitor, click the actions list for the dataflow and then select Run to run the dataflow job.

11.

Click the Refresh Jobs button ( ) to view the latest status of the dataflow job.

You can view the OppRoles dataset after the dataflow job completes successfully.

Note: If you are adding a predicate to a dataset that was previously created, each user must log out and log back in for the predicate to take effect.

225

Analytics Data Integration

Test Row-Level Security for the Dataset

You must verify that the predicate is applied properly and that each user can see the appropriate opportunities.

1. Log in to Wave Analytics as Bill.

2. Open the OppRoles opportunity.

Notice that Bill can’t see his manager Keith’s opportunities anymore. Now, he can see only his opportunity and his subordinate Tony’s opportunity.

EDITIONS

Available in Salesforce

Classic and Lightning

Experience.

Available for an extra cost in

Enterprise, Performance, and Unlimited Editions. Also

Developer