API Changes / Segregation

[jreback]

This will serve as a master issue to record API changes / testing that needs segregation in pandas.

Copy-on-write can help here:

• view propagation: Original numpy array getting altered when changes made to dataframe  pandas-dev/pandas#14953

[wesm]

As a matter of process, the way this should probably work is:

• Segregate unit tests subject to breakage in upstream pandas-dev/pandas
• Rebase pandas2 on upstream branch.
• Write new unit tests that port those "bad" tests to new tests that reflect the changed behavior

what do you think?

[jreback]

yep, should have the test runner exclude them for pandas2 (and add separate tests that only) apply to pandas 2.

so prob need something like

pandas/tests # existing tests
pandas/tests_changing_in_pandas2
pandas/tests_pandas2

with the sub-structure the same as

pandas/tests

but the changing/pandas2 can live on pandas2 repo

[jorisvandenbossche]

From the dev meeting: idea is instead of physically segregating the tests is to leave them at the same place but mark them as "may fail on pandas 2.0"