How to change dataframe column names without changing the values?

Question

I have a bunch of CSV files which are read as dataframes. For each dataframe, I want to change some column names, if a specific column exists in a dataframe:

column_name_update_map = {'aa': 'xx'; 'bb': 'yy'}

In such a map, if 'aa' or 'bb' exists in a dataframe, I want to change the aa to xx, and 'bb' to 'yy'. No values should be changed.

  for file in files:
        print('Current file: ', file)
        df = pd.read_csv(file, sep='\t')
        df = df.replace(np.nan, '', regex=True)
        for index, row in df.iterrows(): 

           pass

I don't think I should use the inner loop, but if I have to do, what's the right way to change the column name only?

score 2 · Accepted Answer · answered Apr 22 '20 at 04:35

2

You can use rename in dataframes

column_name_update_map = {'aa': 'xx', 'bb': 'yy'}
df = df.rename(columns=column_name_update_map)

answered Apr 22 '20 at 04:35

Rajith Thennakoon

3,975
2
14
24

No, 'aa' and 'bb' are only two columns, not all the columns in the df. Would this still work? – Abigail Min NM Apr 22 '20 at 04:40
Abigail your question needs more clarification. Please mention which column names you want to change. If you looking for the unnamed column '' to 'NaN', then you can opt that in the replace() mapping. – Littin Rajan Apr 22 '20 at 04:47
yes,dictionary keys are your old column names and values are new column names. – Rajith Thennakoon Apr 22 '20 at 04:47
I mean if there are 8 columns in the df, and I only need to change 2 of them. Is that still working? – Abigail Min NM Apr 22 '20 at 04:57
It works. You have to opt the parameter 'inplace=True'. Just try the code. – Littin Rajan Apr 22 '20 at 04:59
you can use 'inplace=True' or assign to same dataframe without inplace parameter. – Rajith Thennakoon Apr 22 '20 at 05:00
This one works and I like its simplifity – Abigail Min NM Apr 22 '20 at 05:01

Littin Rajan · Answer 2 · 2020-04-22T05:00:45.977

2

To rename specific columns then follow this code.

Code:

import pandas as pd
import numpy as np

#creating sample dataframe 
df=pd.DataFrame({'aa':[1, 2], 'bb':[3, 4], 'c':[5, 6], '':[7, 8]})

#replace columns 'aa' to 'xx', 'bb'  to 'yy' and '' to 'NaN'
df.rename(columns={'aa':'xx', 'bb':'yy', '':np.nan}, inplace=True)
#display resulting dataframe
print(df)

I hope it would be helpful.

edited Apr 22 '20 at 05:00

answered Apr 22 '20 at 04:56

Littin Rajan

852
1
10
21

replace() got an unexpected keyword argument 'columns' – Abigail Min NM Apr 22 '20 at 04:59
Sorry it was rename() function. – Littin Rajan Apr 22 '20 at 05:01
Does it worked? – Littin Rajan Apr 22 '20 at 05:05

How to change dataframe column names without changing the values?

2 Answers2