Fix Index.difference to avoid collect 'other' to driver side

[itholic]

This PR basically came from SPARK-35683.

This PR fix the wrong behavior of Index.difference in Koalas, based on the comment #1325 (comment) and #1325 (comment)

• it couldn't handle the case properly when self is Index or MultiIndex and other is MultiIndex or Index.

>>> midx1 = ks.MultiIndex.from_tuples([('a', 'x', 1), ('b', 'z', 2), ('k', 'z', 3)])
>>> idx1 = ks.Index([1, 2, 3])
>>> midx1 = ks.MultiIndex.from_tuples([('a', 'x', 1), ('b', 'z', 2), ('k', 'z', 3)])
>>> midx1.difference(idx1)
databricks.koalas.exceptions.PandasNotImplementedError: The method `pd.Index.__iter__()` is not implemented. If you want to collect your data as an NumPy array, use 'to_numpy()' instead.

• it's collecting the all data into the driver side when the other is list-like objects, especially when the other is distributed object such as Series which is very dangerous.

And added the related test cases.

[codecov-commenter]

Codecov Report

Merging #2173 (5f133a5) into master (f971143) will decrease coverage by 0.00%.
The diff coverage is 95.65%.

@@            Coverage Diff             @@
##           master    #2173      +/-   ##
==========================================
- Coverage   95.34%   95.34%   -0.01%     
==========================================
  Files          60       60              
  Lines       13723    13737      +14     
==========================================
+ Hits        13084    13097      +13     
- Misses        639      640       +1     

Impacted Files	Coverage Δ
databricks/koalas/indexes/base.py	97.17% <91.66%> (-0.15%)	⬇️
databricks/koalas/tests/indexes/test_base.py	100.00% <100.00%> (ø)

---

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 f971143...5f133a5. Read the comment docs.