Fix DataFrame.replace to avoid creating a new Spark DataFrame.

[ueshin]

DataFrame.replace() doesn't need to create a new Spark DataFrame.

[ueshin]

also cc @beobest2

[codecov-commenter]

Codecov Report

Merging #1575 into master will decrease coverage by 0.31%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #1575      +/-   ##
==========================================
- Coverage   94.56%   94.24%   -0.32%     
==========================================
  Files          38       38              
  Lines        8773     8668     -105     
==========================================
- Hits         8296     8169     -127     
- Misses        477      499      +22     

Impacted Files	Coverage Δ
databricks/koalas/frame.py	95.94% <100.00%> (-0.89%)	⬇️
databricks/conftest.py	96.22% <0.00%> (-3.78%)	⬇️
databricks/koalas/typedef/typehints.py	85.26% <0.00%> (-2.11%)	⬇️
databricks/koalas/__init__.py	92.15% <0.00%> (-1.97%)	⬇️
databricks/koalas/testing/utils.py	78.98% <0.00%> (-1.81%)	⬇️
databricks/koalas/spark/accessors.py	91.66% <0.00%> (-1.08%)	⬇️
databricks/koalas/internal.py	96.80% <0.00%> (-0.75%)	⬇️
databricks/koalas/mlflow.py	94.87% <0.00%> (-0.13%)	⬇️
databricks/koalas/groupby.py	90.42% <0.00%> (-0.10%)	⬇️
... and 10 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 a02fe6e...abc10dd. Read the comment docs.

[itholic]

LGTM