DOC: Incorrect whatsnew entry for GH6459

[jorisvandenbossche]

The whatsnew entry for PR #6459: http://pandas-docs.github.io/pandas-docs-travis/whatsnew.html at IPython input prompt 22, the example code is:

set_index no longer converts MultiIndexes to an Index of tuples. For example, the old behavior returned an Index in this case (GH6459):

In [22]: df_multi.index
Out[22]: 
MultiIndex(levels=[[u'a', u'b'], [u'c', u'd']],
           labels=[[0, 0, 1, 1], [0, 1, 0, 1]])

# Old behavior, casted MultiIndex to an Index
In [23]: df_multi.set_index(df_multi.index)
Out[23]: 
            0         1
a c  0.903672 -2.201395
  d  0.692797  0.016163
b c -0.315266  0.364278
  d -0.497458  0.531440

[4 rows x 2 columns]

# New behavior
In [24]: df_multi.set_index(df_multi.index)
Out[24]: 
            0         1
a c  0.903672 -2.201395
  d  0.692797  0.016163
b c -0.315266  0.364278
  d -0.497458  0.531440

[4 rows x 2 columns]

The 'old' and 'new' behaviour seems exactly the same, so it not clear what has changed.

[jreback]

hmm maybe the old needs a code block or passing as_tuples=False (which replicates the old)

[jorisvandenbossche]

BTW, there is some code suppressed: https://github.com/pydata/pandas/blob/master/doc/source/v0.14.0.txt#L164 (and I suppose this was meant to create the old/new behaviour)

[jreback]

hmm

I recall this WAS working correctly in the whatsnew maybe something just changed

[jorisvandenbossche]

I suppose it is due to this: #4805 ?
The tuples are converted to a multi-index due to changes to the index constructor?

[jreback]

closed by this: d6241d8

I think was an error when I originally edited the v.0.14.0 notes

you can 'create' the old behavior by passing the internal argument tupleize_cols=False when creating an Index from tuples