Fix from_pandas to handle the same index name as a column name.

[ueshin]

When the input pandas DataFrame has the same index name as a column name, ks.from_pandas() fails with the following error:

>>> pdf = pd.DataFrame({"a": [1, 2, 3], "b": [4, 5, 6]})
>>> pdf.index.name = "a"
>>> pdf
   a  b
a
0  1  4
1  2  5
2  3  6

>>> kdf = ks.from_pandas(pdf)
Traceback (most recent call last):
...
ValueError: cannot insert a, already exists

Resolves #1361, Closes #1375.

[codecov-io]

Codecov Report

Merging #1419 into master will increase coverage by 0.09%.
The diff coverage is 62.50%.

@@            Coverage Diff             @@
##           master    #1419      +/-   ##
==========================================
+ Coverage   95.04%   95.13%   +0.09%     
==========================================
  Files          34       34              
  Lines        7970     7958      -12     
==========================================
- Hits         7575     7571       -4     
+ Misses        395      387       -8     

Impacted Files	Coverage Δ
databricks/koalas/groupby.py	91.76% <0.00%> (+1.30%)	⬆️
databricks/koalas/frame.py	96.61% <100.00%> (ø)
databricks/koalas/internal.py	96.35% <100.00%> (+0.29%)	⬆️

---

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 7b48858...3ea9cfc. Read the comment docs.