Support resetting index with tuple name

[toobaz]

• closes df.reset_index() fails if df.index.name is a tuple #16164
• tests added / passed
• passes git diff master --name-only -- '*.py' | flake8 --diff
• whatsnew entry

Notice the first commit is pure refactoring, and the bug is actually fixed by the one-liner second commit.

[codecov]

Codecov Report

Merging #16165 into master will increase coverage by 0.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master   #16165      +/-   ##
==========================================
+ Coverage   90.87%   90.88%   +0.01%     
==========================================
  Files         162      162              
  Lines       50816    50806      -10     
==========================================
- Hits        46178    46174       -4     
+ Misses       4638     4632       -6

Flag	Coverage Δ
#multiple	88.66% <100%> (ø)	⬆️
#single	40.33% <68.42%> (ø)	⬆️

Impacted Files	Coverage Δ
pandas/core/frame.py	97.58% <100%> (-0.02%)	⬇️
pandas/core/generic.py	91.63% <0%> (+0.32%)	⬆️

---

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 075eca1...e12bca1. Read the comment docs.

[codecov]

Codecov Report

Merging #16165 into master will increase coverage by 0.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master   #16165      +/-   ##
==========================================
+ Coverage   90.85%   90.86%   +0.01%     
==========================================
  Files         162      162              
  Lines       50826    50819       -7     
==========================================
- Hits        46177    46176       -1     
+ Misses       4649     4643       -6

Flag	Coverage Δ
#multiple	88.64% <100%> (+0.01%)	⬆️
#single	40.33% <63.63%> (-0.01%)	⬇️

Impacted Files	Coverage Δ
pandas/core/frame.py	97.58% <100%> (-0.01%)	⬇️
pandas/core/generic.py	91.63% <0%> (+0.32%)	⬆️

---

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 b6f65eb...9e1bdba. Read the comment docs.

[jreback]

this looks good. can you add a test which where you have another level already (and check the resetting), e.g.

In [3]: df = pd.DataFrame([[1, 2]], columns=pd.MultiIndex.from_product([['A'], ['a', 'b']]))
   ...: 
   ...: reind = df.set_index([('A', 'a')], append=True)

it works, just want to cover the bases. otherwise lgtm.

[jorisvandenbossche]

Can you check some more pathological cases?
Eg when the length of the tuple does not match the number of levels? Not sure what the result should be of that (informative error or insert the tuple in first level), though. But good to at least assert its behaviour.

[toobaz]

Not sure what the result should be of that (informative error or insert the tuple in first level), though.

The current error is a quite informative "*** ValueError: Item must have length equal to number of levels.", and I think inserting tuple in first level would break the principle of least surprise.

However, it probably makes sense to use fill_value to fill keys which are shorter than the number of levels, rather than raising, so I added code making this happen.

[jreback]

thanks!

[jorisvandenbossche]

@toobaz Probably a not so important corner case, but could you also check how this interacts with the col_level argument?