CODEWITHSUNDEEP
python-3.xpandasdataframenumpy
Maxi Trien
Maxi Trien

Reputation: 27

Replace class objects with attribute values in dataframe

I have a dataframe in which contains different datatypes, here for example with two classes FirstGroup and LastGroup.

    id1                     id2                     id3...                               
0   (0, FirstGroup.Other)   (0, FirstGroup.Other)   (0, FirstGroup.Static)  ...    
1  (60, LastGroups.Other)  (22, LastGroups.Other)  (49, LastGroups.Static)  ...

All of these objects have data stored that I want to pull out in an efficient way. One example would be: df['id1'].[0][1].name will give me a string variable 'type1'.

I want to have something like this:

    id1            id2            id3...                               
0   (0, 'type1')   (0, 'type1')   (0, 'type2')  ...    
1  (60, 'type1')  (22, 'type1')  (49, 'type2')  ...

I am searching for something similar then the replace method but couldn't fine anything except looping over each row and column.

Upvotes: 1

Views: 286

Answers (1)

mozway
mozway

Reputation: 262149

You can use applymap:

df.applymap(lambda x: (x[0], x[1].name))

full example:

class O:
    def __init__(self):
        self.name = str(id(self))[-5:]
    def __repr__(self):
        return 'CustomObject'

df = pd.DataFrame({'id%s' % i: [(j, O()) for j in range(2)] for i in range(3)})
df.applymap(lambda x: (x[0], x[1].name))

input:

                 id0                id1                id2
0  (0, CustomObject)  (0, CustomObject)  (0, CustomObject)
1  (1, CustomObject)  (1, CustomObject)  (1, CustomObject)

output:

          id0         id1         id2
0  (0, 95984)  (0, 98096)  (0, 97712)
1  (1, 97328)  (1, 98720)  (1, 97280)

Upvotes: 1

Related Questions