-
Notifications
You must be signed in to change notification settings - Fork 55
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Find a solution for multiple BigQuery versions #171
Comments
An issue was created for this problem but it seems there is no activity on it: |
It seems updating version is not trivial because of intertwined dependencies of:
on If we add
which according to this comment is needed by |
Re. the comment you mentioned, there is a background why
So since we now have a version conflict issue, maybe we should consider that |
The One attempt was made in PR 237. The intention was to exclude some scripts for unit tests. But the changes (check Another possible way could be removing |
RE unit test update: I think it may be better to update our unit testing framework to not use Nose as there was a recent thread on the Beam dev site as well about this....We can wait and see what Beam decides to do and follow them (https://issues.apache.org/jira/browse/BEAM-3713). RE BigQuery version: Saman and I discussed this offline, and it may be easier to just use the |
Regardless of whether we change our unit-test framework, we should drop the |
RE google-cloud-storage dependency: I agree. Could we just use Beam's FileSystem instead? It adds a dependency on Beam, but would be simpler to work with and we'd just let Beam deal with the dependency issues. RE Issue #71: Also agree, but that's a separate and more general issue to be resolved in #71. This issue is mainly for the particular BQ dependency. |
Resolved after updating Beam version PR 470. |
In our integration tests, we rely on
google-cloud-bigquery>0.25
, however, Beam needs BigQuerygoogle-cloud-bigquery==0.25
. Even though the two code paths don't currently connect, we should find some way to handle both versions. We currently have an[int_test]
target in our setup.py to handle this case.We can also ask the Beam team whether actually installing bigquery>0.25 is ok (there is a comment that it's only used for tests?).
The text was updated successfully, but these errors were encountered: