Implement DataFrame.dot supporting Series parameter

[xinrong-meng]

Compute the matrix multiplication between the DataFrame and other series only.

        >>> kdf = ks.DataFrame([[0, 1, -2, -1], [1, 1, 1, 1]])
        >>> kser = ks.Series([1, 1, 2, 1])
        >>> kdf.dot(kser)
        0   -4
        1    5
        dtype: int64

        Note how shuffling of the objects does not change the result.

        >>> kser2 = kser.reindex([1, 0, 2, 3])
        >>> kdf.dot(kser2)
        0   -4
        1    5
        dtype: int64

[codecov-io]

Codecov Report

Merging #1945 (6ed71de) into master (901a6f0) will decrease coverage by 0.01%.
The diff coverage is 83.33%.

@@            Coverage Diff             @@
##           master    #1945      +/-   ##
==========================================
- Coverage   94.64%   94.62%   -0.02%     
==========================================
  Files          49       49              
  Lines       10831    10834       +3     
==========================================
+ Hits        10251    10252       +1     
- Misses        580      582       +2     

Impacted Files	Coverage Δ
databricks/koalas/missing/frame.py	100.00% <ø> (ø)
databricks/koalas/frame.py	96.72% <83.33%> (-0.04%)	⬇️
databricks/koalas/series.py	96.85% <0.00%> (-0.12%)	⬇️
databricks/koalas/namespace.py	84.19% <0.00%> (-0.04%)	⬇️
databricks/koalas/missing/indexes.py	100.00% <0.00%> (ø)
databricks/koalas/indexes.py	96.97% <0.00%> (+0.01%)	⬆️
databricks/koalas/internal.py	96.49% <0.00%> (+0.02%)	⬆️

---

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 901a6f0...6ed71de. Read the comment docs.

[itholic]

Seems pandas also support many other types as below, not only Series if they are "aligned".

For example,

>>> pdf
   0  1  2  3
0  0  1 -2 -1
1  1  1  1  1
>>> pser
0    1
1    1
2    2
3    1
dtype: int64

>>> pdf.dot(pser.to_frame())  # DataFrame
   0
0 -4
1  5

>>> pdf.dot(pser.index)  # Index
0   -6
1    6
dtype: int64

>>> pdf.dot([1, 2, 3, 4])  # list
0    -8
1    10
dtype: int64

[xinrong-meng]

Yeah, I should've clarified in the PR description. This PR targets to support other as Series only.

Considering the implementation complexity and user demand, we might want to support other data types later.

Let me modify the PR description and leave a comment for this.

Thanks for looking into this!

[itholic]

Oh, sure I got it :)

[ueshin]

LGTM.

[ueshin]

Thanks! merging.