Tableau Desktop and Web Authoring Help

When you use data blending to combine your data from different data sources, there are a few common issues that might arise. This topic lists them and describes how you can respond to each situation.

Primary and secondary connections are from tables in the same data source

When you drag a field from a secondary data source to the view, you might see a warning that says:

The primary and secondary connections are from tables in the same data source. Instead of linking the connections, use the Data menu to join the data. Joins can integrate data from many tables and may improve performance and filtering.

This warning occurs when the workbook contains separate data sources that connect to the same database. Though you can combine data in this way, Tableau recommends that you use a join or relationship to combine data from the same database instead, as they can leverage some of the database’s native optimizations in ways blending can’t.

Cannot blend the secondary data source because one or more fields use an unsupported aggregation

Data blending has some limitations regarding non-additive aggregates such as COUNTD, MEDIAN, and RAWSQLAGG. Non-additive aggregates are aggregate functions that produce results that cannot be aggregated along a dimension. Instead, the values have to be calculated individually. All Number functions, except for MAX and MIN, are non-additive aggregates. For more information, see Tableau Functions (by Category).

These limitations cause certain fields in the view to become invalid under certain circumstances. If you hover your mouse cursor over one of these invalid fields, you see the following error:

Cannot blend the secondary data source because one or more fields use an unsupported aggregation.

 

This error can occur for one of the following reasons:

• Non-additive aggregates from the primary data source: Non-additive aggregates are only supported in the primary data source if the data in the data source comes from a relational database that allows the use of temporary tables. To work around this issue, consider creating an extract of your data source. Extracts support temporary tables.

• Non-additive aggregates from the secondary data source: Non-additive aggregates are only supported in the secondary data source if the linking field from the primary data source is included in the view and no non-aggregated dimensions from the secondary data source is included in the view. Some number functions can still be used if they include an additive aggregation. For example, use ROUND(SUM([Sales]),1) instead of ROUND([Sales],1).

• Non-additive aggregates from a multi-connection data source that uses a live connection: Multi-connection data sources that connect to data using a live connection do not support temporary tables. Therefore, using a multi-connection data source that connects to data using a live connection prohibits the use of blending functionality with non-additive aggregates. To work around this issue, consider creating an extract of your multi-connection data source. Extracts support temporary tables.

• LOD expressions from the secondary data source: This error can also appear when you use a level of detail expression in a view that uses data blending. To resolve the error, make sure the linking field in the primary data source is in the view before you use an LOD expression from the secondary data source and remove any dimensions, including dimension filters, from the secondary data source.

Data sources that contain logical tables cannot be used as secondary data sources for data blending

The secondary data source can use logical joins as long as the following conditions are satisfied: 

• All active linking fields from the primary data source must be included in the view on a shelf other than the Filters shelf
• All fields from the secondary data source are aggregated, including dimensions

 

Asterisks show in the sheet

When you blend data, make sure that there is only one matching value in the secondary data source for each mark in the primary data source. If there are multiple matching values, you see an asterisk in the view that results after you blend data.

For example, suppose you have two data sources: Population and Superstore. The primary data source, Population, has a field called State. The secondary data source, Superstore, contains fields called State and Segment. In the secondary data source, each state has multiple segments. Alabama, for example, has three segments: Consumer, Corporate, and Home Office.

Primary data source

Secondary data source

When you blend the two data sources on the State field, you create a link where individual state values (in the primary data source) can have multiple segment values (in the secondary data source). In this case, multiple values for segments in the secondary data source for each corresponding state value in the primary data source cause asterisks to appear in the view, as demonstrated by the images below.

Blended data

Resolve asterisks in the sheet

Avoid asterisks in the sheet by making sure that there is only one matching value in the secondary data source for each mark in the primary data source. Though the way you ensure this depends on the data and the view you're trying to create, consider one of the following suggestions to resolve asterisks in the sheet.

• Add field with a higher level of granularity from the primary data source to the sheet.

  For example, suppose you have a view similar to the example used above. Consider adding a field such as City into the sheet because its values are at a higher level of granularity than a field like State.

• Rebuild the view to switch the primary and secondary data sources with each other. In general, you should make the data source whose values have a higher level of granularity the primary data source.

  For example, suppose you have a view similar to the example used above. Consider making Superstore the primary data source and Population the secondary data source.