Change df.columns.names for multiindex columns

Question

I would like to rename one of my levels in a multiindexed columns dataframe in pandas.

df.columns.names

gives me

FrozenList(['level0', 'level1'])

I want to rename 'level0' to 'main'.

I have tried different approaches, none works:

 df.columns.set_names('findingkey', level=0, inplace=True)

gives me TypeError: 'list' object is not callable

I also tried to do it directly:

df.columns.names[0]='main'

with output: TypeError: 'FrozenList' does not support mutable operations.

jezrael · Accepted Answer · 2017-10-04 09:05:32Z

3

Use:

df.columns.names = ['main', 'level1']

Or

df = df.rename_axis(['main', 'level1'], axis=1)

answered Oct 4, 2017 at 9:05

jezrael

868k103 gold badges1.4k silver badges1.3k bronze badges

Sign up to request clarification or add additional context in comments.

5 Comments

First one works, I am just puzzled that df.columns.names[0]='main' did not work then...

Hmm I think error is clear - 'FrozenList' does not support mutable operations, so only assign all new values works.

and there is no method that allows to rename only one?

Unfortunately not :(

only df.columns.names = ['main'] + [df.columns.names[1]] - but it is a bit crazy ;)