-
Notifications
You must be signed in to change notification settings - Fork 87
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
Update type-parser for v2 #1514
Conversation
Codecov Report
|
The two tests that are failing are failing in a dtype merger test:
They use In particular, the latest Azure MacOS runs Hmmm. The test failure is in v1, which has to manually duplicate what NumPy does. v2 probably uses NumPy's own dtype promotion algorithm directly (i.e. concatenate two empty arrays of the appropriate dtype and see what the result is). Actually, it's not even that: it directly asks NumPy to concatenate them: awkward/src/awkward/_v2/contents/numpyarray.py Lines 503 to 505 in 2b6252d
So even if NumPy changes their promotion rules, v2 should be insensitive to that change (it should change along with them). But those actual tests are commented out! awkward/tests/v2/test_0449-merge-many-arrays-in-one-pass.py Lines 53 to 93 in 2b6252d
I should do another PR to uncomment those and make sure they work. I don't care too much about whether v1 only agrees with a range of NumPy versions, it's about to become irrelevant soon. So that PR might drop the v1 test, while enabling the v2 test. |
A more proper fix than #1513. (See continuing issues in scikit-hep/uproot5/pull/620.)