REF: Add ExtensionArray.map

[topper-123]

• closes ExtensionArray.map #23179

Working with the .map method is very uncomfortable in main, because of issues with ExtensionArray:

• ExtensionArray is not required to have a .map method, but may have it.
• Some ExtensionArray subclasses, that do have a map method have one function parameter (mapper), while other have two parameters (mapper and na_action).
• There is fallback functionality for calculation ExtensionArray.map in IndexOps.Mixin._map_values. This means we can calculate .map for ExtensionArrays indirectly, if we wrap them in a Series or Index, but we cannot rely on the ability to do it directly on the ExtensionArray.
• ExtensionArray subclasses often want create their own .map methods
• Because of all the above issues it is often very maze-like to follow code that uses .map

This PR solves this by:

1. moving the functionality in IndexOpsMixin._map_values to a function map_array in algorithms.
2. adding a .map method to ExtensionArray that calls the new map_array function. Subclasses will override this method where needed.
3. Call the new map_array function inside IndexOpsMixin._map_values.
4. Ensuring a common signature for the .map method for ExtensionArray, i.e. always two method parameters (mapper and na_action), so we can reliably call the .map method on any ExtensionArray subclass.

Note that this PR doesn't really add new functionality, because we're mostly just moving code around, though it is now possible to call the map method on any Extensionarray directly, instead of going through e.g. Series.map.

In the main branch, there are currently several ExtensionArray subclasses where na_action='ignore' doesn't work. I'm working on making it work with all of the ExtensionArray subclasses and there is already #51645, which handles Categorical/CategoricalIndex and more will follow.

[jbrockmendel]

Does this close any of #23179, #15706, #32815?

[topper-123]

Does this close any of #23179, #15706, #32815?

• ExtensionArray.map #23179: Yes (adds .map to ExtensionArray)
• Series.map on a Categorical with a function only processes each unique value once #15706: I don't think this issue is solveable while maintaining the performance benefits of a dedicated Categorical.map , (i.e. doing the op only once per category is intended).
• BUG: ExtensionIndex.map doesn't handle na_action correctly #32815: not in this PR, but that is fixed by ENH: add na_action to Categorical.map & CategoricalIndex.map #51645 for the Categorical/CategoricalIndex case. I'll take on DateTimeArray/DateTimeIndex etc. in a follow-up, so that will be fixed shortly.

[jbrockmendel]

copy=False?

[topper-123]

Yeah, good point.

[jbrockmendel]

Might need something in the extension tests?

[topper-123]

Might need something in the extension tests?

I've added some simple tests in the newest version.

[topper-123]

I've updated.

[mroeschke]

The implementation LGTM

[topper-123]

Gentle ping.

I’m intending to follow up with fixes towards #51644, and it would be great if this could be merged.

[mroeschke]

Merge when ready @jbrockmendel

[jbrockmendel]

thanks @topper-123

[jbrockmendel]

@topper-123 does any of your recent work on this close #15706?

[topper-123]

I'vew answered in #15706 to keep the discussion in one place.