API/CoW: Return copies for head and tail #54011
Conversation
There was a problem hiding this comment.
Thanks!
Would there be a usecase for adding a keyword to let you not make the eager copy? That would probably only be useful if you would use head/tail for non-exploration cases, in which case you can use iloc instead (although people might prefer a method in case of method chaining)
doc/source/whatsnew/v2.1.0.rst
Outdated
| @@ -25,6 +25,7 @@ Copy-on-Write improvements | |||
| - The :class:`DataFrame` constructor, when constructing a DataFrame from a dictionary | |||
| of Index objects and specifying ``copy=False``, will now use a lazy copy | |||
| of those Index objects for the columns of the DataFrame (:issue:`52947`) | |||
| - :meth:`DataFrame.head` and :meth:`DataFrame.tail` will now return copies (:issue:`54011`) | |||
There was a problem hiding this comment.
Should we clarify this as "deep copy"? With CoW, the methods already do return an object that behaves as a copy / is a shallow copy.
| @@ -895,16 +895,17 @@ def test_head_tail(method, using_copy_on_write): | |||
| df2._mgr._verify_integrity() | |||
|
|
|||
| if using_copy_on_write: | |||
| assert np.shares_memory(get_array(df2, "a"), get_array(df, "a")) | |||
| assert np.shares_memory(get_array(df2, "b"), get_array(df, "b")) | |||
| assert not np.shares_memory(get_array(df2, "a"), get_array(df, "a")) | |||
There was a problem hiding this comment.
Can you add a comment here to note we explicitly deviate in head/tail methods to do an eager copy (as it now contrasts with most other tests in this file)
|
I would not add a keyword, I am not convinced that there is a proper use-case that would require such a keyword |
doc/source/whatsnew/vX.X.X.rstfile if fixing a bug or adding a new feature.this is something we talked about to avoid messing up references in Jupyter environments (xref #48998)