DOC: sub-section on boolean Index array changes

[jreback]

closes #17740

[jreback]

cc @jorisvandenbossche

[codecov]

Codecov Report

Merging #17803 into master will decrease coverage by 0.01%.
The diff coverage is n/a.

@@ Coverage Diff @@ ## master #17803 +/- ## ========================================== - Coverage 91.24% 91.22% -0.02%  ========================================== Files 163 163 Lines 49971 49971 ========================================== - Hits 45597 45588 -9  - Misses 4374 4383 +9

Flag	Coverage Δ
#multiple	89.03% <ø> (ø)	⬆️
#single	40.24% <ø> (-0.07%)	⬇️

Impacted Files	Coverage Δ
pandas/io/gbq.py	25% <0%> (-58.34%)	⬇️
pandas/core/frame.py	97.74% <0%> (-0.1%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 5653fe4...248830b. Read the comment docs.

[codecov]

Codecov Report

Merging #17803 into master will decrease coverage by 0.02%.
The diff coverage is n/a.

@@ Coverage Diff @@ ## master #17803 +/- ## ========================================== - Coverage 91.25% 91.23% -0.03%  ========================================== Files 163 163 Lines 49971 49971 ========================================== - Hits 45599 45589 -10  - Misses 4372 4382 +10

Flag	Coverage Δ
#multiple	89.03% <ø> (-0.01%)	⬇️
#single	40.24% <ø> (-0.07%)	⬇️

Impacted Files	Coverage Δ
pandas/io/gbq.py	25% <0%> (-58.34%)	⬇️
pandas/core/frame.py	97.74% <0%> (-0.1%)	⬇️
pandas/core/indexes/datetimes.py	95.48% <0%> (-0.1%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 3b4121b...e0ebb7e. Read the comment docs.

[jorisvandenbossche]

Thanks!

I would also add that previously for non-boolean and non-numeric indices, it would raise a KeyError (since the True / False were regarded as labels), and not will do boolean selection.

[jreback]

Iwould also add that previously for non-boolean and non-numeric indices, it would raise a KeyError (since the True / False were regarded as labels), and not will do boolean selection.

not following, what is the example?

[jorisvandenbossche]

So previously Index with True / False values were always regarded as labels, not as a boolean mask. Which resulted in a KeyError for non-numeric indices or actual label-based indexing for numeric/boolean index (the fact that it interpreted True/False as 1/0 is the bug that was been fixed in an earlier PR regarding drop, but the general thing (your recent PR) that boolean index is not seen as labels is the API change you describe in the PR here).

So previous:

In [38]: s = pd.Series([1,2,3], index=['a', 'b', 'c']) In [39]: s.loc[pd.Index([True, False, True])] ... KeyError: "None of [Index([True, False, True], dtype='object')] are in the [index]" In [42]: s = pd.Series([1,2,3]) In [43]: s.loc[pd.Index([True, False, True])] Out[43]: 1 2 0 1 1 2 dtype: int64 

Now:

In [4]: s = pd.Series([1,2,3], index=['a', 'b', 'c']) In [5]: s.loc[pd.Index([True, False, True])] Out[5]: a 1 c 3 dtype: int64 In [6]: s = pd.Series([1,2,3]) In [7]: s.loc[pd.Index([True, False, True])] Out[7]: 0 1 2 3 dtype: int64 

[jorisvandenbossche]

(but it is not perse necessary to add all those example, for me just adding it in the text is fine as well. I suspect that not many people deliberately will have used a boolean index to index a non-numeric index and explicitly catched the KeyError ... The main change is when you actually had a boolean index in the object you are indexing (i.e. the case for which you already have an example))

[jreback]

ic, ok will add that as well.