I tried running a "set column value" action that sets a column value to "Go" on 50 rows in a simple app in desktop Chrome, it took about 10 seconds. This is probably close to the lower bound. There are probably things that could be optimized in there, but if you're seeing it take over a minute then I suspect the difference is related to the number of computed columns in the row and/or having a few particularly expensive AppFormulas. Whenever a row is updated it triggers a recompute of all AppFormulas in the row (as well as things like change-triggered columns or reset conditions).
For those of you seeing such long delays, can you try comparing the time it takes to run it on 50 rows vs the time it takes to run it on 10 rows, and see if the difference is approximately 5x? If it's just a matter of computing a single row being a bit slow due to AppFormulas and then scaling that by 50x, there may not be much we can do to alleviate that other than maybe show a loading spinner. But if the time to update 50 rows is significantly slower than 5 x (time to compute 10 rows), it would suggest that there's something inefficient about the way the bulk action is being applied.
When I try this in my simple app I get 2.3s to run 10 rows, which would imply 11.6s to run 50. So the scaling looks roughly linear when AppFormulas aren't involved.
LinkedIn
Twitter
