Cache an internal pandas object to avoid run twice in Jupyter.

[ueshin]

IPython calls Spark jobs twice due to Notebook calls both __repr__ and _repr_html_.
See ipython/ipython#9771

[codecov-commenter]

Codecov Report

Merging #1564 into master will increase coverage by 0.01%.
The diff coverage is 93.10%.

@@            Coverage Diff             @@
##           master    #1564      +/-   ##
==========================================
+ Coverage   94.51%   94.53%   +0.01%     
==========================================
  Files          38       38              
  Lines        8740     8760      +20     
==========================================
+ Hits         8261     8281      +20     
  Misses        479      479              

Impacted Files	Coverage Δ
databricks/koalas/numpy_compat.py	90.62% <ø> (ø)
databricks/koalas/spark/accessors.py	91.93% <60.00%> (-0.69%)	⬇️
databricks/koalas/base.py	97.37% <100.00%> (+<0.01%)	⬆️
databricks/koalas/frame.py	96.79% <100.00%> (+<0.01%)	⬆️
databricks/koalas/indexes.py	96.85% <100.00%> (+0.03%)	⬆️
databricks/koalas/namespace.py	86.09% <100.00%> (+0.03%)	⬆️
databricks/koalas/series.py	97.57% <100.00%> (+0.12%)	⬆️

---

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 ae57c2a...09af1ae. Read the comment docs.

[itholic]

Seems good to me