Sub-query

Owned by Karen Hogan (Unlicensed)
Last updated: Sept 20, 2022
3 min read

With effect from the pi.2022-03-31 dashboard release:

Sub-query will enable users to introduce charts as filter values, this means that charts are run as sub-queries which will help build the data of the main chart.

Please refer to this page to see a walkthrough of a sub-query filter.

How does Sub-query work:

  • when using an in/not in operator - the first column of the data from the sub-chart is used (multiple values, as described on the tooltip in the screenshot)

  • top/bottom operator - when using a top/ bottom operator, the option to select a chart to filter on will not be available

  • single value operator e.g. Like - the first data point (first row of the first column) is used as the filter value

After selecting what you want to filter on and the operator you want to use you will then be able to select from data or a chart when picking the value to filter on. The type of charts allowed to be used as sub-query are similar to the options that you can select from when using merged charts/ tables. If a chart can be used as a sub-chart of a merged chart or table, it can be used as sub-query. This means that

  • lines, bars, tables are allowed

  • pies, special shapes, maps, merged charts/tables and so on can not be used

The ‘Type to filter’ screen has now been split into two sections - Data and Charts. The Data screen is used in the same way as it was previously (please see this page for more information) and there is now an additional ‘Charts’ screen.

The screenshot below shows the chart types that can be selected from the ‘Charts’ screen. If a chart type can not be used in a sub-query it will not be displayed on the' Charts' screen.

Why use sub-query

By using Sub-query, you will be able to:

  • use sub-chart data from another data connection, e.g. in the following example.

    • chart 1 is from a Sales data connection

    • chart 2 is from a HR data connection

    • chart 1 can be filtered using chart 2

    • this means you could view the sales related to employees who take a lot of holidays

  • use sub-chart data from another chart, which is built with complex sql query structure

Performance with sub-query:

Since sub-query charts are building charts to be used by other charts, the time it takes to display a chart will be the total time it takes to build all the charts used. You can see a breakdown of these times by clicking Chart Tools>Chart Information and selecting the ‘Debug’ tab.

In this example, the total time to build all charts is 586ms. This includes the 352ms taken to build the sub-query chart - you can see this information at the bottom of the debug page on the main chart.

To reduce the length of time taken to draw your chart, try to avoid using lots of sub-query references in your charts.

Potential problems:

  • Cycle referencing: If Chart 1 uses Chart 2, Chart 2 uses Chart 3, Chart 3 uses either Chart 1 or Chart 2 as sub-query, the system shows an error

  • Sub-charts are changed: If a sub-chart is edited afterwards, you may get unexpected results on the main chart while the main chart is not likely to be broken, because the main chart simply just picks the first column or first data point of the sub-charts, it doesn’t care how the sub-charts are configured

  • Sub-charts not having data: again, the main chart is not likely to display data since there isn’t anything relevant