
        
        
            
                
                
                    df.replace(2, "b")
                
            
               A  B
0  b  b
1  3  a

Note that replacement is not done in-place, meaning a new DataFrame is returned and the original df is left intact. We can modify df directly by passing in inplace=True.

Replacing a value in specific columns

To replace 2s in just column A with "b":


        
        
            
                
                
                    df.replace({"A":2}, "b")
                
            
               A  B
0  b  2
1  3  a

Notice how the 2 in column B was not replaced.

Replacing multiple values

To replace multiple values, supply them as a list like so:


        
        
            
                
                
                    df.replace([2,"a"], "b")
                
            
               A  B
0  b  b
1  3  b

Here, we are replacing the values 2 and "a" with "b".

Replacing values based on condition

Consider the following DataFrame:


        
        
            
                
                
                    df = pd.DataFrame({"A":[2,3],"B":[4,5]})
df
                
            
               A  B
0  2  4
1  3  5

To replace values that are greater than 3 with 10:


        
        
            
                
                
                    df[df > 3] = 10
df
                
            
               A  B
0  2  10
1  3  10

Here, we are first creating a boolean mask (DataFrame) where True indicates the values that satisfy the condition:


        
        
            
                
                
                    df > 3
                
            
               A      B
0  False  True
1  False  True

We then fetch a reference to all the entries in df that correspond to True to in the mask:

Finally, performing assignment will update the non-NaN values:


        
        
            
                
                
                    df[df > 3] = 10
df
                
            
               A  B
0  2  10
1  3  10