EA: require size instead of __len__ #28389
Conversation
|
looks like the docs thing is driven by a URLError |
doc/source/whatsnew/v1.0.0.rst
Outdated
| @@ -40,6 +40,7 @@ Other enhancements | |||
| Backwards incompatible API changes | |||
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |||
|
|
|||
| - The :class:`ExtensionArray` interface now requires the author to implement ``size`` instead of ``__len__``. The ``size`` method must _not_ depend on either ``__len__`` or ``shape`` (:issue:`????`) | |||
There was a problem hiding this comment.
you can use this issue number
This PR updates all of our internal (including tests.extension) EAs appropriately, and this is indirectly tested. I've been giving some thought to how to test the "size must not depend on shape or len" requirement. Thinking of something along the lines of a test that monkeypatches |
|
Can we schedule a call sometime this week to talk through this? |
Yes that is a good idea I think. I am -1 on this PR until there is a clear vision and agreement on the next steps. |
|
Wednesday or Thursday after 9 AM Pacific Time work for me |
|
Both work for me (although I would prefer one hour earlier or 1.5 hour later) |
|
these times are ok |
|
An hour earlier on Wednesday works for me. Pending Tom's schedule, let's go for Wednesday at 15:00 UTC. |
|
I should be OK tomorrow (Wednesday) 1 hour earlier (8:00 Pacific / 16:00
CET) or 2.5 hours later (11:30 Pacific / 19:30 CET).
I'll be busy from 12:30 - 1:30 Central.
Free most of Friday.
…On Mon, Sep 16, 2019 at 12:14 PM jbrockmendel ***@***.***> wrote:
Wednesday or Thursday after 9 AM Pacific Time work for me
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#28389?email_source=notifications&email_token=AAKAOIX6R2VCYZMB5WB4TZLQJ65F7A5CNFSM4IVV45DKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD6Z3LYY#issuecomment-531871203>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAKAOIV3YHYKHY5P55WKVCDQJ65F7ANCNFSM4IVV45DA>
.
|
|
did we agree in a date / time?
… On Sep 17, 2019, at 10:55 AM, Tom Augspurger ***@***.***> wrote:
I should be OK tomorrow (Wednesday) 1 hour earlier (8:00 Pacific / 16:00
CET) or 2.5 hours later (11:30 Pacific / 19:30 CET).
I'll be busy from 12:30 - 1:30 Central.
Free most of Friday.
On Mon, Sep 16, 2019 at 12:14 PM jbrockmendel ***@***.***>
wrote:
> Wednesday or Thursday after 9 AM Pacific Time work for me
>
> —
> You are receiving this because you commented.
> Reply to this email directly, view it on GitHub
> <#28389?email_source=notifications&email_token=AAKAOIX6R2VCYZMB5WB4TZLQJ65F7A5CNFSM4IVV45DKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD6Z3LYY#issuecomment-531871203>,
> or mute the thread
> <https://github.com/notifications/unsubscribe-auth/AAKAOIV3YHYKHY5P55WKVCDQJ65F7ANCNFSM4IVV45DA>
> .
>
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.
|
|
I think 15:00 UTC / 11:00 Eastern / 8:00 Pacific. |
|
cc @pandas-dev/pandas-core we're having a call on this in ~25 minutes. https://meet.google.com/gvh-okpg-qux https://docs.google.com/document/d/1k_E_1oSV9VNHgGzepdeCyFdju8ZaXDwI3CvDYkhOtQ8/edit?usp=sharing |
|
Sorry can't hear no sound headphones not working.
…On Wed, 18 Sep 2019, 15:38 Tom Augspurger, ***@***.***> wrote:
cc @pandas-dev/pandas-core
<https://github.com/orgs/pandas-dev/teams/pandas-core> we're having a
call on this in ~25 minutes.
https://meet.google.com/gvh-okpg-qux
https://docs.google.com/document/d/1k_E_1oSV9VNHgGzepdeCyFdju8ZaXDwI3CvDYkhOtQ8/edit?usp=sharing
—
You are receiving this because you are on a team that was mentioned.
Reply to this email directly, view it on GitHub
<#28389?email_source=notifications&email_token=ADEMUXJZ6ZN24S2UBC6ESP3QKI4NJA5CNFSM4IVV45DKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD7AJKLA#issuecomment-532714796>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ADEMUXKA6BQE3XWEWXKMMPTQKI4NJANCNFSM4IVV45DA>
.
|
|
Closing to clear the queue. |
Discussed here
Change the EA interface to require
sizeinstead of__len__.Doing this separately from #27142 makes it much simpler to explain to downstream authors what is changing for them. After this, the rest of the changes necessary for (limited) 2D support should not affect downstream authors.