CODEWITHSUNDEEP
pythonregexpandasfor-loop
Juan C
Juan C

Reputation: 6132

Renaming similar columns from different DataFrames using for loop and Regex in Python

Today I've been working with five DataFrames that are almost the same, but for different courses. They are named df2b2015, df4b2015, df6b2015, df2m2015.

Every one of those DataFrames has a column named prom_lect2b_rbd for df2b2015, prom_lect4b_rbd for df4b2015, and so on.

I want to append those DataFrames, but because every column has a different name, they don't go together. I'm trying to turn every one of those columns into a prom_lect_rbd column, so I can then append them without problem.

Is there a way I can do that with a for loop and regex. Else, is there a way I can do it with other means?

Thanks!

PS: I know some things, like I can turn the columns into what I want using:

re.sub('\d(b|m)','', a)

Where a is the columns name. But I can't find a way to mix that with loops and column renaming.

Edit:

DataFrame(s) look like this:

df2b2015:

rbd   prom_lect2b_rbd
 1          5
 2          6

df4b2015:

rbd   prom_lect4b_rbd
 1          8
 2          9

etc.

Upvotes: 0

Views: 109

Answers (2)

doppler
doppler

Reputation: 997

Something like this, with .filter(regex=)? It does assume there is only one matching column per dataframe but your example permits that.

import pandas as pd
import numpy as np

df1 = pd.DataFrame(np.random.rand(10,3),columns=['prom_lect2b_rbd','foo','bar'])
df2 = pd.DataFrame(np.random.rand(10,3),columns=['prom_lect4b_rbd','foo','bar'])

for df in [df1,df2]:
    colname = df.filter(regex='prom_lect*').columns.format()
    df.rename(columns={colname[0]:'prom_lect_rbd'})

print(df1)
print(df2)

Upvotes: 0

Juan C
Juan C

Reputation: 6132

Managed to do it. Probably not the most Pythonic way, but it does what I wanted:

dfs=[df2b2015,df4b2015,df6b2015,df8b2015,df2m2015]
cols_lect=['prom_lect2b_rbd','prom_lect4b_rbd','prom_lect6b_rbd',
           'prom_lect8b_rbd','prom_lect2m_rbd']

for j,k in zip(dfs,cols_lect):
    j.rename(columns={k:re.sub('\d(b|m)','', k)}, inplace=True)

Upvotes: 1

Related Questions