-
-
Notifications
You must be signed in to change notification settings - Fork 18.4k
DEPR: enforce Series constructor with int dtype deprs #49777
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
Conversation
Thanks @jbrockmendel |
the original test fails on 1.5.x with numpy 1.24.0 - we OK to backport? |
This would enforce a deprecation early in 1.5 so that would be an API change? |
sure, but it's already erroring with numpy 1.24.0 alternatively, do we just pin numpy in the 1.5.x branch and let numpy error? #50384 |
For maximal backwards compat, users should still be able to use pandas 1.5.3 with numpy < 1.24 and not see an API change while being able to use numpy=1.24.0 and not see failures, but that's a lot of effort I presume I would be supportive of pinning numpy in the 1.5.x branch and possibly also including an upper pin for numpy in |
would an upper pin not be pretty limiting, especially on such a widely-used package as numpy, if it's just for the sake of a single test? |
That's fair, this isn't a showstopping issue. If users address the pandas deprecation they shouldn't hit the numpy error correct? If so then I would opt to pin numpy on the 1.5.x build |
I would prefer not to pin numpy as a whole, since pinning tends to break pip's solver. There's also the risk that pip will just pull the version without the pin. |
would the simplest thing be to just conditionally xfail the affected test? |
sure, I've gone for that in #50384 |
doc/source/whatsnew/vX.X.X.rst
file if fixing a bug or adding a new feature.