Unable to export to database issue


I’m having some issues regarding the “Export to database” module and I found this discussion dating back to 2 years now. I’m having the exact same issue, namely no matter which option I chose, the “Create one table per object…” option is highlighted in yellow and after starting CPA with the properties file produced, I get errors whatever i try to do on CPA.

I tried the suggestions in the 2015 discussion but it doesn’t seem to work.

Anyone can help? I’ve attached the pipeline I’m using, a set of images and what my screen shows for the “Export to database” module.



HCS_Ypst_new.cpproj (1.3 MB)



If you set your output directory to a local rather than a network drive, do you have the same “database locked” issue? What about if rather than doing image classification you do object classification?

The yellow is warning you because you likely have different numbers of objects in your different object types; you should set it to “One table per object” instead. From the help (this is from the section on views, but it also applies to single object tables):

An important note is that only objects that are related as primary, secondary or tertiary objects to each other should be combined in a view. This is because the view expects a one-to-one relationship between the combined objects. If you are selecting objects for the view, the module will warn you if they are not related in this way.


Yes, setting the output directory to a local helped, thanks.

Now I have another issue… In my pipeline I flag and skip some images (for example no need to search for cytoplasms if I don’t have a nucleus in the field); this gives me an error message at the end:
"Encountered uncrecoverable error in ExportToDatabase during post-processing:
ufunc ‘isnan’ not supported for the inputs types, and the inputs could not be safely coerced to any supported types according to the casting rules “safe” "

It saves the per_image.csv but not the per_object. It works fine if I hide the “FlagImage” module.

Any ideas?



I’ve done a bunch of things to try to duplicate your issue in another pipeline, but can’t; I suspect it’s one of two things:

  1. It’s a bug in 2.2.0 (which it seems that you’re using) that is fixed in 3.0.0 (which I’m using) OR
  2. It’s something specific in your pipeline; I can’t download it at the moment due to an issue with our forum uploads.

Can you download 3.0.0 and try your pipeline in that to see if it gives the same issue? If it still does, can you upload your pipeline again (either here or to somewhere like GoogleDrive/Dropbox and post a link)? Thanks.