An important point to note about dynamic rows/columns and the Table data format is that the Table can only have one value column (such as one statistic, e.g., unweighted count, mean, etc.). This makes it more difficult to build rules based, for example, on two different statistics. Let’s take a look at the following use case: we want to have dynamic rows sorted by mean satisfaction value for makes, but we want the list to be filtered based on the unweighted count of makes (minimum 1000).
It is not possible to get such a list of makes using just one TabReference. However, we can build two separate TabReferences and combine them to get the desired list of makes.
First, one TabReference could build a dynamic row based on the mean satisfaction, sorted by value in descending order.
We could also build a second TabReference based on the unweighted count, with a filter based on the row value (it will filter out makes with unweighted count < 1000).
It is then possible to integrate the second one as a filter for the first one - we can check if the row label from mean TabReference exists in the count TabReference. In order to do it, we can use a few functions, which are also described in detail in this user guide. We can check the position of (using the function PositionOf) our row label from mean TabReference in the count TabReference. If the result is missing - it means our row does not exist in count TabReference and should be excluded from the dynamic row. Therefore, we can use the IsMissing function to evaluate if this condition is true or not.
Note that since we are nesting here functions in functions, the suggestions mechanism will not be able to correctly suggest a Response label near the end of the expression if you prefer to rewrite it based on the screenshot. As a workaround, you can put the @Response label into the formula before typing the condition and then write the “Not IsMissing…” part before it.
