Fix GroupBy apply, filter, and head to ignore temp columns when ops from different DataFrames.

[ueshin]

Fixing GroupBy.apply, filter, and head to ignore temp columns when ops from different DataFrames, otherwise the result includes the temp columns.

>>> kdf = ks.DataFrame({"a": [1, 2, 3, 4, 5, 6], "b": [1, 1, 2, 3, 5, 8], "c": [1, 4, 9, 16, 25, 36]})
>>> kkey = ks.Series([1, 1, 2, 3, 5, 8])
>>> kdf.groupby(kkey).apply(lambda x: x + x.min())
    a   b   c  __tmp_groupkey_0__
0   2   2   2                   2
5  12  16  72                  16
1   3   2   5                   2
3   8   6  32                   6
2   6   4  18                   4
4  10  10  50                  10

[ueshin]

head has another issue. I'll submit a following PR and add tests there.

[ueshin]

I submitted #1490.

[HyukjinKwon]

no big deal but what about renaming it column_labels_to_exlcude or tmp_column_labels?

[ueshin]

I'd rename it in #1491.

[HyukjinKwon]

I will merge this to unblock but I think we should have a separate class that contained grouping related columns .. it's getting very, very complicated to read ..

[HyukjinKwon]

Sorry I wasn't clear. I agree with the structures and change here. I was talking about a different idea and the fact that we have many grouping related attributes that might have to group up separately.