Changing Data Blending

If your project includes data from two data sources that contain a mixture of attributes and values, and there are match values in one source that don’t exist in the other, then sometimes the system might omit rows of data that you may want to see.

In such cases, you need to specify which source takes precedence over the other for data blending.

For example, we have two data sources (Source A and Source B), which include the following rows. Note that Source A doesn‘t include IN-8 and Source B doesn’t include IN-7.

Description of blending-example-source-tables.gif follows
Description of the illustration blending-example-source-tables.gif

The following results are displayed if the All Rows data blending option is selected for Source A and the Matching Rows data blending option is selected for Source B. Because IN-7 doesn’t exist in Source B, the results contain null Rep and null Bonus.

Description of blending-all-rows-results.gif follows
Description of the illustration blending-all-rows-results.gif

The following results are displayed if the Matching Rows data blending option is selected for Source A and the All Rows data blending option is selected for Source B. Because IN-8 doesn’t exist in Source A, the results contain null Date and null Revenue.

Description of blending-matching-rows-results.gif follows
Description of the illustration blending-matching-rows-results.gif

The visualization for Source A includes Date as an attribute, and Source B includes Rep as an attribute, and the match column is Inv#. Under dimensional rules, these attributes can’t be used with a measure from the opposite table unless the match column is also used.

There are two settings for blending tables that contain both attributes and measures. These are set independently in each visualization based on what columns are used in the visualization. The settings are All Rows and Matching Rows and these describe what rows from a source the system uses when returning data to be visualized.

The system automatically assigns data blending according to the following rules:

  • If a match column is in the visualization, then the sources with the match column are set to All Rows.

  • If an attribute is in the visualization, then its source is set to All Rows and the other sources are set to Matching Rows.

  • If multiple attributes are in the visualization and all come from the same source, then that source is set to All Rows and the other sources are set to Matching Rows.

  • If attributes come from multiple sources, then the source listed first in the project's elements panel is set to All Rows and the other sources are set to Matching Rows.

To change data blending:
  1. Select a visualization on the canvas, and in the visualization toolbar click Menu, then click Properties.
  2. In the Properties dialog, click Data Sets.
  3. In the Data Sets tab, click Auto and then select Custom to view how the system determined blending.
  4. Adjust the blending settings as necessary.
    • At least one source needs to be assigned to All Rows.
    • If both sources are All Rows, then the system assumes that the tables are purely dimensional.
    • You can’t assign both sources to Matching Rows.