Model Builder interface and processing are very slow

2116
4
01-09-2021 12:53 PM
Labels (2)
MichaelD
New Contributor

I am running Arc Pro 2.7.0 and all of my files are being stored locally via connected folders.

I recently built some models, which I then combined into one master model. At first, everything worked fine and the interface was reasonably quick, given the number of operations I asked it to do. But, out of nowhere seemingly, it has become extraordinarily slow. For example, it will take nearly a minute to "think" if I attach a new input to a tool; same for if I change a setting within a tool then go to close the box; same for if I validate or save the model. This extends to running processes as well. When I click run, it will just say "Running...." for at least a minute before finally going to run the tool in its normal manner with details below the progress bar, etc. 

I have attached my Arc file and a PDF of one of the models I'm using, since Python files aren't allowed to be attached. Since my data isn't also attached, I don't know how helpful that will be, however.

Thanks for any help!

4 Replies
DanPatterson
MVP Esteemed Contributor

tool parameter validation perhaps

Model Builder Tool Interface Lag - GeoNet, The Esri Community

see Curstice Price's response


... sort of retired...
MichaelD
New Contributor

Thanks - I'd seen that post before, though, and it only offers one tip; it might help, but my issue happened pretty suddenly even after it was mainly constructed, i.e. I don't think that the slowness is merely due to my model being overly-complex.

0 Kudos
curtvprice
MVP Esteemed Contributor

Note, your aprx does not include your models, for anyone to see it you would share the tbx, or a complete (hopefully stripped down) copy of the project with all its components in a zipped folder or project package.

My only suggestion is to make sure all your models are validated before you call them from another model. Nesting models and using iterators is certainly required when you have a lot of tools. Using iteration can simplify your model (fewer tools to validate)  so @DanPatterson is right on there.

However, even if you are very organized, new inputs (ie ones different from the inputs you used when building the model) may trigger validation of all your tools, which can take a long time for a complex model.  In my experience there is a level of complexity (number of tools) beyond which you want to move to Python to avoid ModelBuilder tool validation.

Hoping the new upcoming toolbox format (2.8?) may provide some faster/smarter validation or maybe more control to optimize validation. We shall see!

DanPatterson
MVP Esteemed Contributor

I don't use modelbuilder, but those multiple Five_Parcel ovals makes me think that you are in need of some "iterator".  Everything is going to be kept in memory until it reaches the very end is my guess.

Since you do a bunch of select by attributes and calculate fields in a sequential fashion rather than an iterative one... it is the I recently built some models, which I then combined into one master model .  If the submodels aren't isolated tasks but use/reuse the same data, then iteration is the route


... sort of retired...