[ARCHIVED] We have obscured the user ids so join keys can be shared without sharing actual Canvas IDs

[rgiovannetti762]

I was trying to use the requests file to collect information on the number of students accessing their course home page over a one week period. I noted that a lot of the activity in the course would be from the instructor, so I attempted to map the user_id in the requests file to the user_id in the user_dim file When I did this I found that the user_id in the requests file is not the id for the actual user that made the request. Is this true or am I missing something?

In the FAQ for canvas data there is a statement:

We have obscured the user ids so join keys can be shared without sharing actual Canvas IDs

Is that what is going on?

[cward]

user_id in any of the tables maps to user_dim.id, so I'm not quite sure what you mean when you say "user_id in the requests file is not the id for the actual user that made the request".

Do you mean you don't find expected users in the requests table?

If that's the case, we found a bug in our underlying technology stack that corrupted user_ids in the requests table. This bug has been patched, and we're in the process of rolling out a corrected requests dataset for everyone.

[rgiovannetti762]

Thanks for follow-up

Yes that is exactly what was happening I was not finding the expected users in the requests table; they did not match the students registered in the courses.

This was my first time using the data and I thought I was doing something wrong or corrupting the data when I retrieved it or linked to it. Just knowing that the data should match is what I was looking for. I will try again and see if my results are matching up better. I was ending up with courses that showed no instructor role or more instructors than were in the course. 

[rgiovannetti762]

Fantastic

That answers that question

Thanks for your assistance with this

Reno