BUG: max_colwidth controls data, not headers

[jreback]

statsmodels/statsmodels#1584

So should do something with the headers (maybe do a '....')

In [2]: df = DataFrame(1,index=range(2),columns=['short','really long'])

In [3]: df
Out[3]: 
   short  really long
0      1            1
1      1            1

[2 rows x 2 columns]

In [6]: with pd.option_context('display.max_colwidth',5):
   ...:     print(df)
   ...:     
   short  really long
0     1      1       
1     1      1       

[2 rows x 2 columns]

[jseabold]

Just a thought and this is more of a related enhancement request, but it'd also be cool (for string columns and indices) to try to be smart when truncating to still give identifiable names. This is what stata does.

prrriiiiiiicccccccceeeeee
prrriiiiiiicccccccceeeeee2

becomes

prrriiiiii~e
prrriiiiii~2 

In truncated outputs.

[jreback]

hmm (as the '...' is pretty common for continuations)

prrriiiiii...e
prrriiiiii...2

[jseabold]

Yeah, sure. I prefer the ellipsis just offering an example.

[mroeschke]

Appears this hasn't gotten traction in a while so closing