Agfdhyk

I am trying to use the replace method several times in order to change the indeces of a given level of a multiindex pandas' dataframe.

As seen here: Pandas: Modify a particular level of Multiindex, @John got a solution that works great so long the replace method is used once.

The problem is, that it does not work if I use this method several times.
E.g.

df.index = df.index.set_levels(df.index.levels[0].str.replace("dataframe_",'').replace("_r",' r'), level=0)

I get the following error message:

AttributeError: 'Index' object has no attribute 'replace'

What am I missing?

Use str.replace twice:

idx = df.index.levels[0].str.replace("dataframe_",'').str.replace("_r",' r')

df.index = df.index.set_levels(idx, level=0)

Another solution is converting to_series and then replace by dictionary:

d = {'dataframe_':'','_r':' r'}

idx = df.index.levels[0].to_series().replace(d)

df.index = df.index.set_levels(idx, level=0)

And solution with map and fillna, if large data and performance is important:

d = {'dataframe_':'','_r':' r'}

s = df.index.levels[0].to_series()

df.index = df.index.set_levels(s.map(d).fillna(s), level=0)

Sample:

df = pd.DataFrame({

        'A':['dataframe_','_r', 'a'],

        'B':[7,8,9],

        'C':[1,3,5],



}).set_index(['A','B'])



print (df)

              C

A          B   

dataframe_ 7  1

_r         8  3

a          9  5



d = {'dataframe_':'','_r':' r'}

idx = df.index.levels[0].to_series().replace(d)

df.index = df.index.set_levels(idx, level=0)

print (df)

      C

A  B   

   7  1

 r 8  3

a  9  5