Hello!
When working with ModelBuilder, tools that I have already run and are in the 'has-been-run' state are very often reset to 'ready-to-run' when I add data to the model or make other changes to the workflow that are not upstream of the change and thus should not have any affect on the reset model elements. It really turns models into a mess because rather than re-running every time a tool resets, I end up adding the data from the previously run tool into the model so I'm not constantly re-running my workflows when this resetting happens. Any help in remedying this would be much appreciated.
Thanks!
Chris
@BrentBrockand @ChrisD do you use spatial joins, join field (or merge-nodes) in your models ... i think i have identified a problem with field mapping at least in those two (or 3) operations -> i.e. a downstream object shows fields in a dropdown that can't exist anymore at that state of the model but have existed in a step before my spatial join and join field operations (i am not sure about merge !) - i guess the attempt of field-list-matching downstream triggers an upstream-request and the latter is setting back each of the upstream objects
I use all of the above. I put everything in models and use modelbuilder all the time for table manipulation, vector analysis, raster analysis, etc. If you have a specific model that triggers your issue, definitely reach out to tech support and share it with them. Hasn't resulted in a fix yet, but presumably the more examples and tickets the better. Take care.
Hi,
We can repro a case but want to make sure it is exactly what you see. Possible to share just the model or a snapshot of it?
Thanks!
The current model I'm working on is purely raster so no joins of any kind, but I've had this issue in every model I've built (well over a hundred by now) since migrating to ArcGIS Pro several years ago. I opened a support case in June and the issue was assigned #BUG-000150197 but I've seen no indication on progress.
I can also confirm that the problem with setting 'ready-to-run' state works both ways as making significant changes to process tools does not always reset the state. Yesterday I added a tool to the process chain but when I executed the model, it ignored the major modification I made and only processed the last couple steps in the model well downstream of where modifications were made.
Clearly there are major issues with the way tool states are being managed in Pro. This has been a major headache for me as I'm working on a model covering most of North America at 30m resolution. Even on my beast of a workstation, unnecessarily re-running tools adds hours and often days waiting for results. The problem is compounded by the fact that some tools like Raster Calculator ignore cancel requests so if a model incorrectly resumes processing on one of these tools, I'm left with the equally bad choices of waiting an hour for the tool to run before the model finally cancels, or killing the whole ArcGIS process in Task Manager.
Is there a way to upvote bugs like we can for feature requests?
Hello all,
We re investigating right now and based on that we will prioritize it immediately. The tools rerunning unnecessarily is definitely non needed.
Fantastic! Thanks!
Thank you!👍
unfortunately, the reset to "ready-to-run" error still exists in the current patch - which is pretty annoying if you run models where some nodes take hours to calculate !!!
Hi Chris,
Possible to share the model or a snapshot of it please?
Thanks in advance!
I opened a ticket for the issue – case #03263553. I spent some time on the phone with an analyst trying to reproduce the issue, and wasn’t able to in the few minutes we were in the screensharing session. This is the second time I’ve talked to an analyst and not been able to reproduce the issue during the session. I mean, it happens very frequently for me, but not all the time for every new tool connection. Very frustrating when it doesn't happen during the support call. They always ask me to send in a model with all the underlying data when I do actually have the issue occur, but I haven't taken the time to do it yet. So, feel free to reach out to tech support and reference this case and share models with issues. Such a weird issue that obviously happens a lot for some of us, but presumably not for everyone.