Revert "DEPR: is_copy arg of take (#30615)"

[jorisvandenbossche]

This reverts commit 7796be6.

See #27357 (comment) and #30615 (comment)

[jreback]

-1 on this. the goal is to remove is_copy kwarg

i would rather fix take later if necessary, rather that redo this PR and fix take later with necessary

this is a step backwards and doesn’t accomplish anything

[jorisvandenbossche]

i would rather fix take later if necessary, rather that redo this PR and fix take later with necessary

If we want to fix take later, why should we remove that behaviour of take now which we want to preserve later? (the fact that take is an actual copy)

BTW, I can also try to do a PR that actually fixes take. This is basically #30784, but in addition we would need to keep a private version of take that does the set_is_copy thing and ensure to use that private version where appropriate (i.e. mainly in the indexing code).
But then I first like to have some agreement on that general idea before putting time in it.

[jreback]

i would rather fix take later if necessary, rather that redo this PR and fix take later with necessary

If we want to fix take later, why should we remove that behaviour of take now which we want to preserve later? (the fact that take is an actual copy)

BTW, I can also try to do a PR that actually fixes take. This is basically #30784, but in addition we would need to keep a private version of take that does the set_is_copy thing and ensure to use that private version where appropriate (i.e. mainly in the indexing code).
But then I first like to have some agreement on that general idea before putting time in it.

that would be preferable

when private and public take were consolidated a while back this must have been broken

also include the tests for no setting with copy warning

thanks

[jorisvandenbossche]

@jreback ok, did what is described above in #30784

So closing this PR then