CODEWITHSUNDEEP
pythonpandas
Rakesh Adhikesavan
Rakesh Adhikesavan

Reputation: 12826

How to replace values in a dataFrame by checking for a condition?

I have the following data frame:

A  |   Date1  | Date2 
10 | 2/2/2016 | 3/2/2016
11 | 1/5/2016 | 1/5/2016
12 | 2/3/2016 | 2/3/2016
13 | 1/5/2016 | 3/2/2013

If the value in Date1 is equal to Date2, I want to make the value in column A as 0. End result:

A  |   Date1  | Date2 
10 | 2/2/2016 | 3/1/2016
0  | 1/5/2016 | 1/5/2016
0  | 2/3/2016 | 2/3/2016
13 | 1/5/2016 | 3/2/2013

I want to do this without writing a for loop. Can I use apply?

You can recreate my df:

df = pd.DataFrame([[10, "2/2/2016", "3/2/2016" ] , [11, "1/5/2016", "1/5/2016"] , [12 , "2/3/2016" , "2/3/2016" ] , [13, "1/5/2016", "3/2/2013"]])
df.columns = ['A','B','C']

Upvotes: 1

Views: 59

Answers (2)

piRSquared
piRSquared

Reputation: 294258

Using jezrael's setup

Setup

import pandas as pd


df = pd.DataFrame([[10, "2/2/2016", "3/2/2016" ] ,
                   [11, "1/5/2016", "1/5/2016"] ,
                   [12 , "2/3/2016" , "2/3/2016" ] ,
                   [13, "1/5/2016", "3/2/2013"]])
df.columns = ['A','B','C']

using loc

df.loc[df.B == df.C, 'A'] = 0

print df

    A         B         C
0  10  2/2/2016  3/2/2016
1   0  1/5/2016  1/5/2016
2   0  2/3/2016  2/3/2016
3  13  1/5/2016  3/2/2013

Upvotes: 1

jezrael
jezrael

Reputation: 862641

Use mask:

import pandas as pd


df = pd.DataFrame([[10, "2/2/2016", "3/2/2016" ] ,
                   [11, "1/5/2016", "1/5/2016"] ,
                   [12 , "2/3/2016" , "2/3/2016" ] ,
                   [13, "1/5/2016", "3/2/2013"]])
df.columns = ['A','B','C']
print (df)
    A         B         C
0  10  2/2/2016  3/2/2016
1  11  1/5/2016  1/5/2016
2  12  2/3/2016  2/3/2016
3  13  1/5/2016  3/2/2013

df['A'] = df.mask(df.B == df.C, 0)
print (df)
    A         B         C
0  10  2/2/2016  3/2/2016
1   0  1/5/2016  1/5/2016
2   0  2/3/2016  2/3/2016
3  13  1/5/2016  3/2/2013

Solution with where:

df['A'] = df.where(df.B != df.C, 0)
print (df)
    A         B         C
0  10  2/2/2016  3/2/2016
1   0  1/5/2016  1/5/2016
2   0  2/3/2016  2/3/2016
3  13  1/5/2016  3/2/2013

Upvotes: 3

Related Questions