PathLikeSelector.make_index() could be sped up by relying upon DataFrame.from_tuples()

[lebedov]

Since pandas.DataFrame.from_tuples() inserts NaN to fill in index rows for tuples that contain fewer elements than other tuples, we could rewrite make_index() to not explicitly extract levels from the expanded selector by relying upon from_tuples(); this might speed up the method. The presence of NaN values in the index shouldn't cause any problems.

[lebedov]

Nikul noticed that the presence of NaNs in the index breaks selection with the DataFrame.ix attribute. We can still use from_tuples() when all of the rows are the same length, though.