cancel
Showing results for 
Search instead for 
Did you mean: 
Highlighted
Surveyor

SIS import batch mode deletes courses/enrollments we include in CSV files

Hi all, I'm wondering if anyone has had the same experience as us with the Batch mode for SIS imports. It's described here: https://canvas.instructure.com/doc/api/file.sis_csv.html. We're using multi-term batch mode, where we send a file of active terms as well as a courses file and an enrollment file.

The idea is that any SIS-created course or enrollment (in an active term) that is no longer in the CSV files we send, will be deleted from Canvas. This was running fine for 2 weeks and allowing us to keep our SIS and Canvas precisely in sync (which was a challenge previously)

However, we've had 3 instances where the Batch import stops working as advertised. It appears to not "see" a block of records that we send (as active) in the CSV, and deletes a swath of courses or enrollments. I've had to turn off the Batch flag on our imports while we resolve the issue. There's nothing different about the files we sent or the API call we use when it doesn't work.

I opened a Canvas support case, but it's taking a while and I wanted to see if anyone has experienced the same issue. I'd also love to hear from anyone who uses the Batch import and finds that it works well and does what they expect, to help understand if we're an outlier here.

Thanks,

Ted

Labels (3)
0 Kudos
3 Replies
Highlighted
Community Advocate
Community Advocate

It happened to us too. We had an issue with our sis imports getting "stuck" and not processing. The subsequent uploads deleted enrollments that were in the enrollments.csv file. I asked for an explanation but never received one. The uploads just started working again the next day on their own with no code changes on our part. 

Highlighted

Thanks, @MattHanes -- would you be open to sharing the size and frequency of your batch imports? We have about 15,000 records we send every 5 minutes

Highlighted

@tbularzik sorry I missed your question. We send 300,000 records about 3 times a day in our biggest import file. We have some smaller ones that we send as well but that's the biggest one.