Use select_column if possible when read_hdf

[dzubo]

If in pd.read_hdf(..., columns=['col_name], ...) we have only one column, then it would be much faster to use store.select_column() instead of store.column().

pandas/pandas/io/pytables.py

Line 368 in fa231e8

return store.select(key, auto_close=auto_close, **kwargs)

Any caveats on implementing this?

[gfyoung]

Not that I know of. Go for it!