When attempting to resize a column in a table, if you have selections turned on, single or multiple, it can lead to a null
selection in the selectedRow
/selectedRows
values. If I had to speculate on the cause, it is because it is selecting the header row, but I'm not entirely sure about this. Something weird is happening here, and this is a massive bug.
Goal:
When resizing a column on a table with selections enabled (single or multiple - doesn't matter which), it should not reset (or in the case of multiple, add) the selection to null
.
Recreation Steps:
Create a table, and set the selection
type to multiple or single. From here, open the debugger to State, and view the selectedRow
/selectedRows
for the table. There are multiple routes you can take to replicate this:
- Resize a column, and observe
selectedRow
/selectedRows
for each column being set tonull
. - Select a(n) row(s), then resize a column, and observe
selectedRow
/selectedRows
for each column being set tonull
.
Details:
I am on Retool version 3.28.3. It's almost as if the header 'row' is being selected when a column is resized, and this should not be the case.
Sources
Here is a video replicating this issue using both single and multiple-selection methodologies:
Single Selection:
Multiple Selection
Notice: There are two selections, then a null
selection in this screenshot below, this can go on. When you resize a column, it will push in one of these null values. So you could have null, value, value. Or value, null, value, value. There are many ways this could be replicated:
JSON Export
Here is an export of an example application I created to showcase this issue, although this is quite as simple as creating a project, adding a table, resizing a column, and observing the results of the effect on selectedRow
/selectedRows
.
Table Selection.json (119.1 KB)