DEPR: NDFrame.set_axis inplace defaults to false #27525

[ghost]

pandas/pandas/core/generic.py

Lines 594 to 598 in 76247c1

	.. warning::
	``inplace=None`` currently falls back to to True, but in a
	future version, will default to False. Use inplace=True
	explicitly rather than relying on the default.

has been there since #20164 #16994 (issue was #14636), part of 0.21.0.

With discussion of plans to deprecate similar functionality from set_index in #24046, it's time to make sure set_axis conforms to the rest of pandas in this.

[WillAyd]

looks pretty good minor comments

[WillAyd]

In case anyone is looking for it I think you meant to reference #16994 (the OP just points to a documentation update)

[WillAyd]

Minor whatsnew quibbles but lgtm otherwise

[ghost]

Thanks for the review

[jreback]

thanks @pilkibun

[ghost]

cc @toobaz, you had some sentiments relating this to the set_index vs. set_axis debate in #16994 (comment) in relation to set_index vs. set_axis. I'm not what your position was, but It sounds like two years ago you were against both the little-endians and the big-endians ;)

... Can we haz some docs now?

[toobaz]

@pilkibun I think I'm missing something, as I don't understand.

But related to the comment you link, I guess it is probably time to remove this warning I had introduced...

pandas/pandas/core/generic.py

Line 650 in 37f29e5

'set_axis now takes "labels" as first argument, and '

(and the subsequent reassignment).

[ghost]

@pilkibun I think I'm missing something, as I don't understand.

You objected to "publicizing" set_axis in the docs because of its defects and also that changing the default to inplace=False (this PR) is necessary before it can become officially recommended.

Now that that's done, I hope you'll push for making a decision on set_index vs. set_axis, so that we can finally document what the "official" method is.

[toobaz]

OK, now I see where you're going.

The defects of set_axis are, as you saw, a thing of the past, and they are/were entirely unrelated to set_index. You can just go on with a nice docs PR "publicizing" set_axis if you like.

The API contains a lot of "official" methods to do a lot of different things.

Please stop provoking.