Skip to content

Dask-friendly nan check in xr.corr() and xr.cov() #5284

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

Merged
merged 10 commits into from
May 27, 2021
Merged

Dask-friendly nan check in xr.corr() and xr.cov() #5284

Changes from 1 commit
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
b8d6daf
initial changes
AndrewILWilliams May 9, 2021
227b622
Using map_blocks to lazily mask input arrays, following #4559
AndrewILWilliams May 16, 2021
6619a6c
Adding lazy corr cov test with `raise_if_dask_computes`
AndrewILWilliams May 26, 2021
637ab80
adding test for one da without nans
AndrewILWilliams May 26, 2021
fff4556
checking ordering of arrays doesnt matter
AndrewILWilliams May 26, 2021
2756488
adjust inputs to test
AndrewILWilliams May 26, 2021
58c441d
add test for array with no missing values
AndrewILWilliams May 26, 2021
a9bd723
added whatsnew
AndrewILWilliams May 27, 2021
8f50b09
Merge branch 'master' into dask_corr_cov
AndrewILWilliams May 27, 2021
ccb5f67
fixing format issues
AndrewILWilliams May 27, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 17 additions & 1 deletion xarray/core/computation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1327,7 +1327,23 @@ def _cov_corr(da_a, da_b, dim=None, ddof=0, method=None):
# 2. Ignore the nans
valid_values = da_a.notnull() & da_b.notnull()

if not valid_values.all():
def _nan_check(d):
if d.all():
return True
else:
return False

if is_duck_dask_array(valid_values.data):
# assign to copy - else the check is not triggered
_are_there_nans = valid_values.copy(
data=valid_values.data.map_blocks(_nan_check, dtype=valid_values.dtype),
deep=False,
)

else:
_are_there_nans = _nan_check(valid_values.data)

if not _are_there_nans:
Copy link
Contributor

@dcherian dcherian May 10, 2021 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks @AndrewWilliams3142 .

Unfortunately I don't think you can avoid this for dask arrays. Fundamentally bool(dask_array) will call compute. So this line if not _are_there_nans will call compute.

I think the only solution here is to special case this optimization for non-dask DataArray, and always call .where(valid_values) for dask arrays but maybe someone else has an idea.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

OK I guess you could stick the

valid_values = ...
if valid_values.any():
    ...

in a helper function and map_blocks that.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hmm, ok. Silly question but does it matter that _are_there_nans is a scalar reduction of valid_values? So valid_values is still left uncomputed. Would this help the bottleneck at all?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No because the program needs to know valid_blocks.any() to decide which branch of the if-else statement to execute. So it gets computed.

da_a = da_a.where(valid_values)
da_b = da_b.where(valid_values)

Expand Down