-
Notifications
You must be signed in to change notification settings - Fork 928
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
New array conversion methods #9236
New array conversion methods #9236
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I enjoyed reading this PR, great consistency, clarity, and structure. PRs on this team are looking sharp! Just a typo, a couple of docs recommendations, and a question.
b71316b
to
ac7c43d
Compare
rerun tests |
Codecov Report
@@ Coverage Diff @@
## branch-21.12 #9236 +/- ##
===============================================
Coverage ? 10.75%
===============================================
Files ? 116
Lines ? 19429
Branches ? 0
===============================================
Hits ? 2090
Misses ? 17339
Partials ? 0 Continue to review full report at Codecov.
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Mostly looks good with 2 Qs.
rerun tests |
@gpucibot merge |
This PR fixes breaking changes made in: rapidsai/cudf#9236 Authors: - GALI PREM SAGAR (https://github.com/galipremsagar) Approvers: - Jordan Jacobelli (https://github.com/Ethyling) - Ajay Thorve (https://github.com/AjayThorve) URL: #348
This PR adds the
to_numpy
andto_cupy
methods.to_numpy
is the preferred method for generating numpy arrays as of pandas 1.0 when older methods liketo_matrix
were removed. This PR deprecates those old methods as well as their gpu counterparts (such asas_gpu_matrix
) and instead addsto_cupy
as the preferred method of getting a__cuda_array_interface__
compliant array view. The new methods are also preferred to the.values
and.values_host
accessors, which are not yet deprecated in pandas but are likely to be deprecated at some point due to the ambiguity of their copy semantics.