Pandas: HDFStore method to get the columns from a table

[cpcloud]

I have some frames that have 44 ish columns with names I can't really remember. I think it might be useful to be able to do something like

store.get_index('hdf5/table', axis=1)

which would return the columns of hdf5/table. @jreback thoughts?

[cpcloud]

Right now I'm doing store.get_storer(key).non_index_axes[0][1].

[jreback]

hmm

that looks good

only caveat is that axis 1 is just a trivial meta data read while axis 0 actually has to select_column (and user might by except this) but as long as it's doc ed guess it's ok

[cpcloud]

eh i'll just rename my columns to something more meaningful .. this probably not worith it.

[jreback]

actually I think this is a nice idea

let's leave open

[Karel-van-de-Plassche]

This is also a nice feature for large datasets, as one doesn't need to lead the whole table in memory just to view column names if one uses the method @cpcloud describes.