Use iloc to make sure it retrieves the first element.

[ueshin]

Use iloc in infer_pd_series_spark_type to make sure it retrieves the first element.

Before:

>>> infer_pd_series_spark_type(pd.Series(['a','b','c'], index=[10,20,30]))
Traceback (most recent call last):
...
KeyError: 0

After:

>>> infer_pd_series_spark_type(pd.Series(['a','b','c'], index=[10,20,30]))
StringType

Note that currently the function in only called from InternalFrame.from_pandas which always resets index before calling it.

[codecov-io]

Codecov Report

Merging #2037 (9b8c3d7) into master (88212f3) will decrease coverage by 0.06%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #2037      +/-   ##
==========================================
- Coverage   94.64%   94.57%   -0.07%     
==========================================
  Files          54       54              
  Lines       11402    11294     -108     
==========================================
- Hits        10791    10681     -110     
- Misses        611      613       +2     

Impacted Files	Coverage Δ
databricks/koalas/typedef/typehints.py	93.80% <100.00%> (ø)
databricks/koalas/__init__.py	86.56% <0.00%> (-4.48%)	⬇️
databricks/koalas/testing/utils.py	79.89% <0.00%> (-1.33%)	⬇️
databricks/koalas/generic.py	92.39% <0.00%> (-0.35%)	⬇️
databricks/koalas/plot/core.py	92.24% <0.00%> (-0.29%)	⬇️
databricks/koalas/namespace.py	84.17% <0.00%> (-0.24%)	⬇️
...ricks/koalas/tests/plot/test_series_plot_plotly.py	97.93% <0.00%> (-0.13%)	⬇️
databricks/koalas/mlflow.py	95.12% <0.00%> (-0.12%)	⬇️
databricks/koalas/indexing.py	92.53% <0.00%> (-0.10%)	⬇️
databricks/koalas/indexes/multi.py	95.72% <0.00%> (-0.08%)	⬇️
... and 16 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 88212f3...9b8c3d7. Read the comment docs.

[itholic]

Nice