CODEWITHSUNDEEP
pythonpandasdataframe
dpacman
dpacman

Reputation: 3899

How to encode a pandas row with custom values?

I would like to encode the row values of the selected column in pandas.

For example, let's assume I have the following dataframe

    Col_A  Col_B
0    SV     NT
1    NT     P
2    SV     I
3    P      SV
4    I      P

I have a dictionary containing the encode values for each row of a column like below

encode_values = { 
    "Col_A" : {"NT": 1, "SV": 0, "P": 1, "I": 0},
    "Col_B": {"NT": 10, "SV": 0, "P": 10, "I": 0}
}

I want a new dataframe with the custom encodings for each of those columns

Result: -

    Col_A  Col_B
0    0     10
1    1     10
2    0     0
3    1     0
4    0     10

Upvotes: 1

Views: 768

Answers (2)

Poe Dator
Poe Dator

Reputation: 4903

Try replace with dictionary as parameter, apply separately to each series in the df:

df.Col_A = df.Col_A.replace (encode_values['Col_A'])
df.Col_B = df.Col_B.replace (encode_values['Col_B'])

Upvotes: 0

Henry Yik
Henry Yik

Reputation: 22503

You can apply and map using column names:

print (df.apply(lambda d: d.map(encode_values[d.name])))

   Col_A  Col_B
0      0     10
1      1     10
2      0      0
3      1      0
4      0     10

Upvotes: 2

Related Questions