bug fix when all columns match but no rows match

[fdosani]

Fixes #276

@SimonBFrank would you mind pulling down this branch and seeing if it fixes your issue. Long story short, there are no results and the subsequent dictionary with values which get displayed out is not able to generate because of None values.

More of a hack since we will be deprecating this legacy Spark implementation (#275) for a much more readable and logically similar to Pandas version. Just catching the TypeError for this situation and putting out {} for columns_with_any_diffs and columns_fully_matching

[SimonBFrank]

When I compare the two dataframes with join columns ["id", "label"] this is the result:

Dataframe 1:

id	label	tmp
1	foo	1
2	bar	1

Dataframe 2:

id	label	tmp
3	foo	1
4	bar	1

****** Column Summary ******
Number of columns in common with matching schemas: 3
Number of columns in common with schema differences: 0
Number of columns in base but not compare: 0
Number of columns in compare but not base: 0

****** Row Summary ******
Number of rows in common: 0
Number of rows in base but not compare: 2
Number of rows in compare but not base: 2
Number of duplicate rows found in base: 0
Number of duplicate rows found in compare: 0

****** Row Comparison ******
Number of rows with some columns unequal: 0
Number of rows with all columns equal: 0

****** Column Comparison ******
Number of columns compared with some values unequal: 0
Number of columns compared with all values equal: 0

****** Columns with Unequal Values ******
Base Column Name  Compare Column Name  Base Dtype     Compare Dtype  # Matches  # Mismatches
----------------  -------------------  -------------  -------------  ---------  ------------

I believe Number of rows with some columns unequal and Number of columns compared with some values unequal should be 2 since only the values in the column id are different. Additionally, Columns with Unequal Values should have id.

[fdosani]

I believe Number of rows with some columns unequal and Number of columns compared with some values unequal should be 2 since only the values in the column id are different. Additionally, Columns with Unequal Values should have id.

I might not be following right, but I think since none of the join columns match (["id", "label"]) in this situation it should all be 0. It is joining on both the fields, not just one.

[SimonBFrank]

I believe Number of rows with some columns unequal and Number of columns compared with some values unequal should be 2 since only the values in the column id are different. Additionally, Columns with Unequal Values should have id.

I might not be following right, but I think since none of the join columns match (["id", "label"]) in this situation it should all be 0. It is joining on both the fields, not just one.

Whoops, I must've been late and I didn't understand it correctly. LGTM