Thanks to visit codestin.com
Credit goes to github.com

Skip to content

np.datetime('NaT') == np.datetime('NaT', 'us') is no longer True on NumPy 1.10 #6452

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.

Sign up for GitHub

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

Jump to bottom
Closed
shoyer opened this issue Oct 12, 2015 · 1 comment
Closed

np.datetime('NaT') == np.datetime('NaT', 'us') is no longer True on NumPy 1.10 #6452

shoyer opened this issue Oct 12, 2015 · 1 comment

Comments

@shoyer
Copy link
Member

shoyer commented Oct 12, 2015

I'm not quite sure exactly what happened, but whereas np.datetime64('NaT') == np.datetime64('NaT', 'us') is True on NumPy 1.9, it's False on NumPy 1.10.

I tested this out on a few NumPy/Python versions, and in some cases this comparison results in warnings. In particular:

  • Python 2.7, NumPy 1.9: True / No warning
  • Python 3.4, NumPy 1.9: True / DeprecationWarning: Implicitly casting between incompatible kinds. In a future numpy release, this will raise an error. Use casting="unsafe" if this is intentional.
  • Python 2.7, NumPy 1.10: False / DeprecationWarning: elementwise == comparison failed; this will raise an error in the future.
  • Python 3.4, NumPy 1.10: False / No warning

Based on git blame, it looks like the latest deprecation warning introduced by @njsmith in #5964.

xref pydata/xarray#618
@shoyer
Copy link
Member Author

shoyer commented Oct 12, 2015

This seems like a bug in the datetime64 casting rules for generic time units. I'm working on a fix.

@shoyer shoyer mentioned this issue Oct 13, 2015
Merged
@ssanderson ssanderson mentioned this issue Dec 7, 2015
Closed
ssanderson pushed a commit to quantopian/zipline that referenced this issue Jan 22, 2016
MAINT: Fix warnings raised by numpy 1.10.
951cca0 
Fixes lots of ambiguous calls to `np.full` where the result value was
float but the input was an integer, all of which were throwing
FutureWarnings.

Also works around numpy/numpy#6452 by using
NaT values with explicit units.
ssanderson pushed a commit to quantopian/zipline that referenced this issue Jan 22, 2016
MAINT: Fix warnings raised by numpy 1.10.
909f02c 
Fixes lots of ambiguous calls to `np.full` where the result value was
float but the input was an integer, all of which were throwing
FutureWarnings.

Also works around numpy/numpy#6452 by using
NaT values with explicit units.
shoyer added a commit to shoyer/numpy that referenced this issue May 24, 2018
@shoyer
TST: regression test for numpygh-6452
2666831
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@shoyer