Dear all,
When I try to use the tool "Train Random Trees Classifier" with a composite of 36 layers, it gives me the error message "Error 999999: Error executing function. Failed to execute (TrainRandomTreesClassifier)". I tried a lot of things to find out why this error occurs, and now it seems to me that the tool simply cannot manage more than 32 input layers. Has anyone experienced the same and does anyone know how to overcome this problem?
I'm thankful for any help!
does it work with part of the inputs?
everything in the same coordinate system?
spaces in folder paths?
resources? is it failing because you lack computing resources
ArcMap or PRO?
Thanks for your answer!
-yes,
-yes,
-no,
-I don't think so? How do I know this?
-ArcMap.
http://pro.arcgis.com/en/pro-app/tool-reference/spatial-analyst/train-random-trees-classifier.htm
step it back one at a time, until it works, ArcMap is limited in the amount of ram it can use, so try it in PRO if you have a beefy machine ram-wise
I'm not sure if I understand. My computer has all the system requirements needed to work with ArcMap. My computer's RAM is only 8 GB, so it is sufficient, but not enormous.
As Dan mentioned, ArcMap for being a 32 bit application can only access less than 4 GB or RAM (regardless of the amount you have installed), hence the recommendation to try and use Pro (64 bits)
Ok now I get it, thanks a lot to both of you. I will try using Pro then.
I'm running arcMap 10.5.1 and here was my fix to the ERR999999 with Train Random Trees
Open a NEW arcmap session (clean-slate) and try running the tool
I tried everything - Using different training samples (created both in/outside of the training manager), different ancillary multiband rasters (or none), segmented image (both all three bands and just 1 band), changing the parameters, etc.
Running the tool outside the mxd I was performing all my processing in, which had ~100 geoprocessing results, was the fix. Map Documents (especially documents with lots of processing history) are notorious for having "gremlins" in them
Thank you for your answer! I just tried it this way. But unfortunately, it didn't work, either.