Have seen a couple of runs this weekend where the Peloton workflow has attempted to add the same item to the project twice in the same query. GitHub handles this but our workflow cannot handle it downstream. I was unable to replicate the problem this morning. If it materialises again we might consider adding something after this line:
|
issues_discussions = pd.concat( |
|
[issues.data_frame, discussions.data_frame, accessibility.data_frame], |
|
) |
The issues_discussions dataframe uses the ID of the issue/pr/discussion as the index. We can enforce this to be unique (it is certainly expected to be).
Have seen a couple of runs this weekend where the Peloton workflow has attempted to add the same item to the project twice in the same query. GitHub handles this but our workflow cannot handle it downstream. I was unable to replicate the problem this morning. If it materialises again we might consider adding something after this line:
.github/peloton/update_project.py
Lines 1331 to 1333 in 40368fc
The
issues_discussionsdataframe uses the ID of the issue/pr/discussion as the index. We can enforce this to be unique (it is certainly expected to be).