[BUG] TimedeltaIndex intersects with decreasing TimedeltaIndex

[makbigc]

• closes BUG: TimedeltaIndex.intersection fails for decreasing TimedeltaIndex #17391
• tests added (two tests taken from BUG: TimedeltaIndex.intersection #22114)
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

Beside the bug of decreasing TimedeltaIndex, this PR also solve the following

In [2]: idx1 = pd.to_timedelta(range(1,20), unit='s')                           

In [3]: idx2 = pd.to_timedelta((1,2,3,4,5,6,7,8,9,10,20), unit='s')             

In [4]: result = idx1.intersection(idx2)                                        

In [5]: result                                                                  
Out[5]: 
TimedeltaIndex(['00:00:01', '00:00:02', '00:00:03', '00:00:04', '00:00:05',
                '00:00:06', '00:00:07', '00:00:08', '00:00:09', '00:00:10',
                '00:00:11', '00:00:12', '00:00:13', '00:00:14', '00:00:15',
                '00:00:16', '00:00:17', '00:00:18', '00:00:19'],
               dtype='timedelta64[ns]', freq=None)

The above is also added into the tests.

[codecov]

Codecov Report

Merging #25661 into master will increase coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master   #25661      +/-   ##
==========================================
+ Coverage   91.29%   91.29%   +<.01%     
==========================================
  Files         173      173              
  Lines       52961    52966       +5     
==========================================
+ Hits        48350    48357       +7     
+ Misses       4611     4609       -2

Flag	Coverage Δ
#multiple	89.87% <100%> (ø)	⬆️
#single	41.73% <0%> (-0.01%)	⬇️

Impacted Files	Coverage Δ
pandas/core/indexes/timedeltas.py	90.95% <100%> (+0.68%)	⬆️

---

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 c4fa3c9...a7e6853. Read the comment docs.

[codecov]

Codecov Report

Merging #25661 into master will decrease coverage by 0.52%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master   #25661      +/-   ##
==========================================
- Coverage   91.82%   91.29%   -0.53%     
==========================================
  Files         175      173       -2     
  Lines       52551    52966     +415     
==========================================
+ Hits        48256    48357     +101     
- Misses       4295     4609     +314

Flag	Coverage Δ
#multiple	89.87% <100%> (-0.51%)	⬇️
#single	41.73% <0%> (+0.87%)	⬆️

Impacted Files	Coverage Δ
pandas/core/indexes/timedeltas.py	90.95% <100%> (+0.68%)	⬆️
pandas/io/gbq.py	25% <0%> (-62.5%)	⬇️
pandas/io/common.py	72.86% <0%> (-19.01%)	⬇️
pandas/compat/__init__.py	58.03% <0%> (-17.8%)	⬇️
pandas/io/excel/_util.py	78.82% <0%> (-8.68%)	⬇️
pandas/compat/chainmap.py	61.9% <0%> (-4.77%)	⬇️
pandas/core/groupby/categorical.py	95.45% <0%> (-4.55%)	⬇️
pandas/core/computation/common.py	85.71% <0%> (-3.76%)	⬇️
pandas/core/dtypes/cast.py	88.16% <0%> (-3.2%)	⬇️
pandas/io/s3.py	86.36% <0%> (-3.12%)	⬇️
... and 118 more

---

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 a1fee91...90d32f2. Read the comment docs.

[TomAugspurger]

Capitalize bug. Reference the method :meth:`TimedeltaIndex.intersection`.

Remove the trailing whitespace.

[TomAugspurger]

Sorting can be relatively expensive. Are we able to avoid this?

I think this would also break sort=False.

[TomAugspurger]

I would think that we could extract the i8 values and re-use Int64Index.sort. Is that an option?

[jreback]

can you merge master