Use of Replace() in Python Dataframe for Multiple Columns but same value

Question

Query: I need to replace the 1 old value with the 1 new value for a bunch of columns (not all columns) in a dataframe. The question is about the syntax to be used. Is there a shorter syntax?

Sample Dataframe:

df = pd.DataFrame({'A': [0,1,2,3,4],
                   'B': [5,6,7,0,9],
                   'C': [2,0,9,3,0],
                   'D': [1,3,0,5,2]})

I need all 0 to be replaced with 10 in the above df but only for column A and C (Not for B or D).

Code that I use to do this:

Method 1: Two separate commands.

df['A'].replace({0:10},inplace=True)
df['C'].replace({0:10},inplace=True)

Method 2: One command using dictionary in dictionary

df.replace({'A': {0:10}, 'C': {0:10}},inplace=True)

Method 3: Keeping new value out of dictionary

df.replace({'A':0,'C':0},10,inplace=True)

Expected Outcome:

    A  B   C  D
0  10  5   2  1
1   1  6  10  3
2   2  7   9  0
3   3  0   3  5
4   4  9  10  2

I am able to get expected outcome using all three methods. But I have a doubt that can we give a list of columns and enter old and new values for replacement only once?

Something like:

df.replace({['col_ref'...]:{'old':'new'})
#OR
df['col_ref'...].replace()

In my scenario, there are 26 columns out of 52 that need replacing, and the value is to be replaced through a regex command. Now I can store the regex command as a variable and use the method 2 to do this. But this also requires entering the variable name for 26 times. Is there any shorter way where I can enter these 26 columns and the regex replacement {'r':'r2'} only once?

tylerjames · Accepted Answer · 2022-04-23 20:12:58Z

2

I was looking on how to do this quicker myself this week and found this method and setup to handle instead of a for loop:

col_list = ['A', 'B']
df[col_list] = df[col_list.replace(0,10,inplace=True)

If you are using regex for a string:

col_list = ['A', 'B']
df[col_list] = df[col_list.replace('[\$,]','',regex=True, inplace=True)

answered Apr 23, 2022 at 20:12

tylerjames

1339 bronze badges

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

Comments

Andreas · Accepted Answer · 2025-03-04 15:59:19Z

1

cols = ["A", "B"]

df[cols] = df[cols].replace(0, 10).values

Prevent exceptions and warnings such as ValueError: Columns must be same length as key by using .values:

answered Mar 4 at 15:59

Andreas

9,2854 gold badges20 silver badges47 bronze badges

Comments

Veki · Accepted Answer · 2022-01-17 17:07:11Z

0

I tried this.

for col in [list of columns]:
    df.replace({col:{'r':'r2'}},regex=True,inplace=True)

This is the shortest way I could think of to write minimum code characters.

However, if there is a faster way, other answers are welcome.

answered Jan 17, 2022 at 17:07

Veki

5816 silver badges23 bronze badges

Collectives™ on Stack Overflow

Use of Replace() in Python Dataframe for Multiple Columns but same value

3 Answers 3

Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related