Series.reset_index does not consider 'name' parameter

[isilopekel]

System information

• OS Platform and Distribution: Linux Ubuntu 16.04
• Modin version : 0.7.3
• Python version: 3.6.7
• Code we can use to reproduce:

Create dummy.csv with following content: 
a,b,c,
1,2,3
4,5,6

if __name__ == "__main__":
    import modin.pandas as mpd
    df=mpd.read_csv("dummy.csv", sep=",")
    df_count = df.groupby(["a", "b"]).size().reset_index(name="Count")
    print(df_count.columns)

    import pandas as pd
    df=pd.read_csv("dummy.csv", sep=",")
    df_count = df.groupby(["a", "b"]).size().reset_index(name="Count")
    print(df_count.columns)

Describe the problem

Multi-column grouping loses the column names. Output of the code snippet:
Index(['level_0', 'level_1', 'Count'], dtype='object')
Index(['a', 'b', 'Count'], dtype='object')

[devin-petersohn]

Hi @isilopekel thanks for the report!

This issue is due to groupby.size not being implemented. We have a start on this, it will be in the next release.

Normally we would return a modin.pandas.Series, but this will return a pandas.Series object because it is very old code from before the Series implementation.

Link to groupby.size implementation:

modin/modin/pandas/groupby.py

Lines 365 to 366 in 985adb0

	def size(self):
	return pandas.Series({k: len(v) for k, v in self._index_grouped.items()})

[dchigarev]

It seems that the problem here is not in groupby at all, but in reset_index. It seems that it does not consider custom name

if __name__ == "__main__":
    import pandas
    import modin.pandas as pd
    
    md_sr, pd_sr = pd.Series([1, 2, 3, 4]), pandas.Series([1, 2, 3, 4])

    print(md_sr.reset_index(name="Custom name").columns) # Index(['index', 0], dtype='object')
    print(pd_sr.reset_index(name="Custom name").columns) # Index(['index', 'Custom name'], dtype='object')

[devin-petersohn]

@dchigarev Great find! Please change the title of the issue to reflect the correct issue.