BUG: Passing original properties of DataFrame and Series subclasses to their constructors

[YankoFelipe]

• closes BUG: Subclassed DataFrame doesn't persist _metadata properties across binary operations #34177 BUG: Subclassed DataFrame doesn't persist _metadata properties across binary operations
• Tests added and passed if fixing a bug or adding a new feature
• All code checks passed.
• Added type annotations to new arguments/methods/functions.
  No new arguments, methods nor functions.
• Added an entry in the latest doc/source/whatsnew/vX.X.X.rst file if fixing a bug or adding a new feature.

While implementing a subclass of Dataframe, I found that in some operations, objects that subclass's DataFrame and Series forget their original properties.

import pandas as pd

class SubclassedSeries(pd.Series):
    _metadata = ['original_property']
    def __init__(self, data=None, original_property=None, *args, **kwargs):
        super().__init__(data, *args, **kwargs)
        self.original_property = original_property
    @property
    def _constructor(self):
        return SubclassedSeries
    @property
    def _constructor_expanddim(self):
        return SubclassedDataFrame


class SubclassedDataFrame(pd.DataFrame):
    _metadata = ['original_property']
    def __init__(self, data=None, original_property=None, *args, **kwargs):
        super().__init__(data, *args, **kwargs)
        self.original_property = original_property
    @property
    def _constructor(self):
        return SubclassedDataFrame
    @property
    def _constructor_sliced(self):
        return SubclassedSeries

## __init__
df = SubclassedDataFrame({'lkey': ['foo', 'bar', 'baz', 'foo'], 'value': [1, 2, 3, 5]}, original_property='original_property')
print(f'__init__: {df.original_property}')

## Select
select_df = df[df['value'] == '1']
print(f'Select: {df.original_property}')

## loc
loc_df = df.loc[df['lkey'] == 'foo']
print(f'loc: {loc_df.original_property}')

## Concat
df = pd.concat([df, df])
print(f'Concat: {df.original_property}')

## Merge
df1 = SubclassedDataFrame({'lkey': ['foo', 'bar', 'baz', 'foo'],
                   'value': [1, 2, 3, 5]}, original_property='original_property')
df2 = SubclassedDataFrame({'rkey': ['foo', 'bar', 'baz', 'foo'],
                   'value': [5, 6, 7, 8]}, original_property='original_property')
merged_df = df1.merge(df2, left_on='lkey', right_on='rkey')
print(f'Merge: {merged_df.original_property}')

## Series
series = df['value']
print(f'Series: {series.original_property}')

## Sum
sum_df = df1 + df2
print(f'Sum: {sum_df.original_property}')

(df * 2).added_property

This functionality is critical for my project so I decided to fix it. I had the following considerations:

• __init__ parameters in both parent classes weren't modified.
• _name stays in Series._metadata and their children because @540db96b. It is removed right before the calling the constructor and added again in __init__ of parent class.
• I added a unit test covering the cases mentioned above. I mentioned BUG: Subclassed DataFrame doesn't persist _metadata properties across binary operations #34177 but I think _metadata items of subclassed pd.Series are not propagated into corresponding SubclassedDataFrame #32860, QST: Subclassing and concat #35415, API: _constructor and subclasses #32638, Transfer _metadata from Subclassed DataFrame to Subclassed Series #19850 and DOC/API: document how to use metadata #8572 are related.
• Each DataFrame child must come with their respective Series child (and vice versa). Both children must contain and handle the same _metadata parameters.
• The constructors' parameters of both children must begin with data=None. The example with (self, data=None, **_metadata, *args, **kwargs) was the cleanest I was able to think without having to clean *args and **kwargs (like the old unit test in pandas/tests/frame/test_subclass.py:MySubclassWithMetadata). Not enforcing data as the first parameter and having to clean up *args and **kwargs felt as wrong as adding *args and **kwargs to __init__ in both parents.

My environment consist on SUSE 15.6 and Python 3.11.11

Edit: Removed WIP comments about tests and code checks.

[YankoFelipe]

Any chance to review this PR anytime soon?
@noatamir @TomAugspurger @jorisvandenbossche

[github-actions]

This pull request is stale because it has been open for thirty days with no activity. Please update and respond to this comment if you're still interested in working on this.

[mroeschke]

Thanks for the pull request, but it appears to have gone stale. If interested in continuing, please merge in the main branch, address any review comments and/or failing tests, and we can reopen.