BUG: sefault in concat of CategoricalIndex

[chris-b1]

• closes .concat crashes Python #16111
• tests added / passed
• passes git diff upstream/master --name-only -- '*.py' | flake8 --diff
• whatsnew entry

[codecov]

Codecov Report

Merging #16133 into master will decrease coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master   #16133      +/-   ##
==========================================
- Coverage   90.86%   90.86%   -0.01%     
==========================================
  Files         159      159              
  Lines       50771    50773       +2     
==========================================
+ Hits        46133    46134       +1     
- Misses       4638     4639       +1

Flag	Coverage Δ
#multiple	88.64% <100%> (-0.01%)	⬇️
#single	40.33% <25%> (ø)	⬆️

Impacted Files	Coverage Δ
pandas/core/indexes/category.py	98.44% <100%> (ø)	⬆️
pandas/core/common.py	90.68% <0%> (-0.35%)	⬇️

---

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 8d122e6...5b6be78. Read the comment docs.

[codecov]

Codecov Report

Merging #16133 into master will decrease coverage by 0.02%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master   #16133      +/-   ##
==========================================
- Coverage   90.86%   90.83%   -0.03%     
==========================================
  Files         159      159              
  Lines       50771    50773       +2     
==========================================
- Hits        46133    46122      -11     
- Misses       4638     4651      +13

Flag	Coverage Δ
#multiple	88.61% <100%> (-0.03%)	⬇️
#single	40.33% <25%> (ø)	⬆️

Impacted Files	Coverage Δ
pandas/core/indexes/category.py	98.44% <100%> (ø)	⬆️
pandas/plotting/_converter.py	63.54% <0%> (-1.82%)	⬇️
pandas/core/common.py	90.68% <0%> (-0.35%)	⬇️

---

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 8d122e6...5b6be78. Read the comment docs.

[TomAugspurger]

@chris-b1 your test example doesn't segfault for me on master, but the fix does work. Do the share the same underlying issue?

[chris-b1]

Yeah, the underlying problem is the we end up with a bad indexer, which is passed to take_nd, which doesn't bounds check, and end up with either a bad result (like the test) or a segfault. Here is one that segfaults on master.

categories = list('vabcdefghijklmnopqrstu')

a = pd.Series(1, index=pd.CategoricalIndex(categories[:5], categories=categories))
b = pd.Series(2, index=pd.CategoricalIndex(categories[3:10], categories=categories))
c = pd.Series(3, index=pd.CategoricalIndex(categories[8:15], categories=categories))

result = pd.concat([a, b, c], axis=1)

[jreback]

lgtm merge away