ENH: chunks=None #106

Activity

1. alimanfoo

   added this to the v2.2 milestone on Jan 24, 2017
2. jakirkham commented on Jan 24, 2017

   

   jakirkham

   on Jan 24, 2017

   Last edited by jakirkham

   Member

   More actions

   Would this make it a change in current behavior as it is already None?

   Edit: Would it make sense to have the current behavior go through True instead. ( https://github.com/alimanfoo/zarr/issues/107 )
3. alimanfoo commented on Jan 24, 2017

   

   alimanfoo

   on Jan 24, 2017 via email

   MemberAuthor

   More actions

   I think I want to avoid changing the current behaviour. I haven't thought this through thoroughly, but I was wondering about setting the default to chunks=True, which means auto-chunking (which is the current behaviour). Then if the user explicitly states chunks=None that means chunk shape equals array shape.

   …

   On Tuesday, January 24, 2017, jakirkham ***@***.***> wrote: Would this make it a change in current behavior as it is already None? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <https://github.com/alimanfoo/zarr/issues/106#issuecomment-274890875>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAq8Qu_BnNpsjPdkAOhyQDi3Csqx60kiks5rVkHFgaJpZM4Lsdvx> .

   -- Alistair Miles Head of Epidemiological Informatics Centre for Genomics and Global Health <http://cggh.org> The Wellcome Trust Centre for Human Genetics Roosevelt Drive Oxford OX3 7BN United Kingdom Email: alimanfoo@googlemail.com Web: http://purl.org/net/aliman Twitter: https://twitter.com/alimanfoo Tel: +44 (0)1865 287721
4. jakirkham commented on Jan 24, 2017

   

   jakirkham

   on Jan 24, 2017

   Member

   More actions

   To be fair, I do think this proposal results in a better default.

   Just thinking of the right way to deprecate this. Normally adding a warning about chunks=None would be ok, but it is the default too. So that could get really noisy. Maybe using a decorator that checks for chunks being explicitly set or doing some introspection of the arguments passed in.
5. alimanfoo commented on Jan 25, 2017

   

   alimanfoo

   on Jan 25, 2017 via email

   MemberAuthor

   More actions

   FWIW I think it would be OK to just change the default to chunks=True and include some documentation about the change in the docstrings and release notes. I don't think anyone will have been explicitly passing chunks=None before now. The behaviour when no chunks argument is specified by the user (autochunking) will be unchanged.

   …

   On Tuesday, January 24, 2017, jakirkham ***@***.***> wrote: To be fair, I do think this proposal results in a better default. Just thinking of the right way to deprecate this. Normally adding a warning about chunks=None would be ok, but it is the default too. So that could get really noisy. Maybe using a decorator that checks for chunks being explicitly set or doing some introspection of the arguments passed in. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <https://github.com/alimanfoo/zarr/issues/106#issuecomment-274901798>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAq8QjdRz4b4tXCDJbb5ebfDanMwKFRUks5rVkqigaJpZM4Lsdvx> .

   -- Alistair Miles Head of Epidemiological Informatics Centre for Genomics and Global Health <http://cggh.org> The Wellcome Trust Centre for Human Genetics Roosevelt Drive Oxford OX3 7BN United Kingdom Email: alimanfoo@googlemail.com Web: http://purl.org/net/aliman Twitter: https://twitter.com/alimanfoo Tel: +44 (0)1865 287721

   👍React with 👍1Reacted by jakirkham
6. alimanfoo

   added

   to do

   on Oct 31, 2017
7. alimanfoo

   self-assigned this

   on Nov 14, 2017
8. alimanfoo

   added

   in progressSomeone is currently working on thisSomeone is currently working on this

   and removed

   to do

   on Nov 14, 2017
9. alimanfoo

   closed this as completedin #183on Nov 14, 2017
10. alimanfoo

    added

    enhancementNew features or improvementsNew features or improvements

    release notes doneAutomatically applied to PRs which have release notes.Automatically applied to PRs which have release notes.

    and removed

    in progressSomeone is currently working on thisSomeone is currently working on this

    on Nov 14, 2017