Format change in backend spreadsheet

[BigCat402]

I have an app where my employees go out and inspect our infrastructure. It has quite a few columns that can be edited. I do have a ChangeCounter column however the employee that has to enter the info into a separate database has to go through and determine which cell has changed in that row. Is there anyway to highlight the cell in the backend spreadsheet if it has been changed? I couldn't find a conditional formatting option that would work. Or is there a better way that I could determine which cell has been edited?

[dbaum · Accepted solution]

You could inspect the spreadsheet's edit history.

Within the app, you could log more details about the change, although it could be tedious to set up. For instance:

• Create a separate ChangeCounter column for each column you want to monitor.
• Create a column with an expression (maybe an initial value expression with reset on edit enabled) that compares before and after values for each column and returns a list of the names of changed columns.

[BigCat402]

dbaum,

What do you mean about creating a column with an expression that could return the list of names of changed columns? Can you give me an example?

[dbaum]

Here's where I would start experimenting: Access column values before and after an update - AppSheet Help. For example:

LIST(
IFS([[_THISROWBEFORE].[Column1] <> [Column1], "Column1", 
IFS([[_THISROWBEFORE].[Column2] <> [Column2], "Column2", 
...
) - 
LIST("")

I haven't tested that technique, and the timing and sequence of the various steps in updating a row may be such that it doesn't work for your purpose. In that case, you could consider instead using an action to log the desired information--maybe by using a grouped action to first save a copy of the row before opening the form to edit, or maybe just by using an action as the form view's form saved event.

[BigCat402]

dbaum,

Thanks for the help. I think my best option will be creating a ChangeCounter for each column as I really don't need to know the before values.

[dbaum]

That's certainly a valid approach.

Also, to be clear, the expression I drafted for you wouldn't log the before values, but rather only the names of the columns whose before values differed.