BUG: Fix of handle missing CSV MI column names

[gfyoung]

De-hackifies this hack:

pandas/pandas/io/parsers.py

Lines 3206 to 3211 in d78bd7a

	# hack
	if (isinstance(index_names[0], compat.string_types) and
	'Unnamed' in index_names[0]):
	index_names[0] = None
	return index_names, columns, index_col

Setup:

from pandas.compat import StringIO
from pandas import read_csv

data = ",,col\na,c,1\na,d,2\nb,c,3\nb,d,4"
print(read_csv(StringIO(data), index_col=[0, 1]))

data = "NotReallyUnnamed,Unnamed: 0,col\na,c,1\na,d,2\nb,c,3\nb,d,4"
print(read_csv(StringIO(data), index_col=[0, 1]))

Before:

# Why is only `index_names[0]` replaced with `None`?
              col
  Unnamed: 1
a c             1
  d             2
b c             3
  d             4

# Having "Unnamed" in the name doesn't make it replace-able ?
# (this is also surfacing the `index_names[0]` bug too)
              col
  Unnamed: 0
a c             1
  d             2
b c             3
  d             4

After:

# All placeholder names get dropped.
              col
a c             1
  d             2
b c             3
  d             4

# Non-placeholder names never get dropped.
                             col
NotReallyUnnamed Unnamed: 0
a                c             1
                 d             2
b                c             3
                 d             4

[pep8speaks]

Hello @gfyoung! Thanks for submitting the PR.

• There are no PEP8 issues in the file pandas/io/parsers.py !

• There are no PEP8 issues in the file pandas/tests/io/parser/index_col.py !

[codecov]

Codecov Report

Merging #23484 into master will increase coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master   #23484      +/-   ##
==========================================
+ Coverage   92.23%   92.23%   +<.01%     
==========================================
  Files         161      161              
  Lines       51197    51204       +7     
==========================================
+ Hits        47220    47227       +7     
  Misses       3977     3977

Flag	Coverage Δ
#multiple	90.61% <100%> (ø)	⬆️
#single	42.27% <64.7%> (ø)	⬆️

Impacted Files	Coverage Δ
pandas/io/parsers.py	95.62% <100%> (+0.01%)	⬆️

---

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 24ab22f...e18fef9. Read the comment docs.

[gfyoung]

@jreback : I simplified the logic a little bit, but I still needed unnamed_count in some cases because it's computed on a per-iteration basis of a for-loop, whereas unnamed_cols is a global collection. Everything is still green though. PTAL.

[jreback]

it seems like you could just add unamed_cols.add(name) at line 774 (e.g. after the if name == '' ?

[gfyoung]

Not quite. You need to add the name that you get post-mangling (e.g. if there are dupes). That's why you have to "keep track" after the logic ending at 794.

[jreback]

thanks!