DataFrame.to_string() remove extra white space between columns

[lodagro]

DataFrame.to_string() uses one extra white space between columns for sign alignment.
Remove it, for user who need exact column width, how to do this nicely?

see also mailing list discussion

[wesm]

this is still a problem post PR, will have a look

In [1]: paste
In [331]: DF=DataFrame({'i':[1,1,1],'j':[1,2,3],'k':[4,4,4],'mass_g_d':
[1.1,2.5,10.4]})

In [332]:

In [333]: intformatter = lambda x: '%10i' %x

In [334]: floatformatter = lambda x: '%10.5f' %x

In [335]: print(DF.to_string(columns=['k','i','j','mass_g_d'],
   .....:                    index_names=False,
   .....:                    colSpace=0,
   .....:                    formatters={'k':intformatter,
   .....:                                'i':intformatter,
   .....:                                'j':intformatter,
   .....:                                'mass_g_d':floatformatter}))
## -- End pasted text --
pandas/core/frame.py:939: FutureWarning: colSpace is deprecated, use col_space
  FutureWarning)
  k          i          j          mass_g_d  
0          4          1          1    1.10000
1          4          1          2    2.50000
2          4          1          3   10.40000

[wesm]

Sorry, my bad, after looking at the formatters, I think this is a right-justify problem. Hmmmmm

[wesm]

OK, kludge my way to victory, and only took 6 minutes:

In [335]: print(DF.to_string(columns=['k','i','j','mass_g_d'],
   .....:                    index_names=False,
   .....:                    colSpace=0, justify='right',
   .....:                    formatters={'k':intformatter,
   .....:                                'i':intformatter,
   .....:                                'j':intformatter,
   .....:                                'mass_g_d':floatformatter}))
## -- End pasted text --
pandas/core/frame.py:939: FutureWarning: colSpace is deprecated, use col_space
  FutureWarning)
           k          i          j   mass_g_d
0          4          1          1    1.10000
1          4          1          2    2.50000
2          4          1          3   10.40000

[wesm]

What would ya'll think about having justify='right' as the default options? Looks more like R, but maybe it would make life easier overall. Eyeballed what hierarchical columns and it seems fine to me

In [16]: print frame.T.ix[:, :6].to_string(justify='right')
first     foo                     bar             baz
second    one     two   three     one     two     two
exp                                                  
A      -1.137 -1.6775  0.7792  0.3529  0.6557 -0.1332
B      -0.568 -0.7236  0.6556 -0.2667 -0.8273  1.1148
C       0.890 -0.4917 -0.3609  0.5157 -0.2967  1.1162