CODEWITHSUNDEEP
pythonpandas
FooBar
FooBar

Reputation: 16508

Renaming stacked columns

I have a column set as

>>> test2.columns
MultiIndex(levels=[[u't070199', u't070299', u't070201', u't070105', u't070104', u'employment'], [u'foo', u'fubar']],
           labels=[[0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5], [0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1]],
           names=[u'foo', u'status'])

I would like to create a "mesh grid" out of the two levels, I am doing the following for now:

level0 = test2.columns.levels[0]
level1 = test2.columns.levels[1]
columnNames = []

for l in level0:
    for j in level1:
        columnNames.append(l+'_'+j)
columnNames
['t070199_foo', 't070199_fubar', 't070299_foo', 't070299_fubar', ...]

This is exactly what I want, but it doesn't look clean at all. Taking the two sets, I suppose there is a Python way of intertwining these two more comfortably. Or, perhaps, a natural way that pandas provides to create "one level column names" out of "2 level column names".

Upvotes: 1

Views: 76

Answers (2)

FooBar
FooBar

Reputation: 16508

test2.columns.ravel() gives me the following tuple: ('t070199', 'foo'), ('t070199', 'fubar')....

So, I can do 

test2.columns = ['_'.join(x) for x in test2.columns.ravel()]

which gives me exactly what I wanted.

Upvotes: 0

Julien Spronck
Julien Spronck

Reputation: 15433

One thing you can do to make it a little nicer is using a list comprehension:

columnNames = [lev0 + '_' + lev1 for lev0 in level0 for lev1 in level1]

Upvotes: 2

Related Questions