I would like to submit a feature request related to the Table object Column Selector, which I find to be a very powerful and useful feature.
Today, the Column Selector requires absolute column indexes using functions like CreateList(). Because the number of visible month columns changes throughout the year, the index of the final column also changes. For example:
- In February, the table may contain columns [ Region | Jan | Feb | Monthly Average ], so CreateList(1,4) works.
- In March, the table may contain Region | Jan | Feb | Mar | Monthly Average ], so CreateList(1,4) now incorrectly returns the March column instead of the Monthly Average.
To work around this, I often have to create a separate hidden table that only contains the Monthly Average column, which can add complexity to report development and maintenance. A very helpful enhancement would be support for reverse column indexing in the Column Selector.
- For example: CreateList(1, -1)
Where -1 would represent the rightmost column, evaluated from right to left. This would allow developers to reliably include trailing summary columns such as averages, totals, or variances, regardless of how many dynamic columns precede them. The same would be beneficial for tables where the column dimension can be changed by a variable before the user downloads the report. For example, breaking out by Region may return 4 columns, while breaking out by Customer could return 40 columns.
I feel this functionality would significantly simplify report design for dynamic tables and reduce the need for creating one-off workaround objects to achieve the desired report.