Instantiate a class in a groupby method

Question

I have the following class for a file:

class File:
    def __init__(self, file_name, md5):
        self.file_name = file_name
        self.md5 = md5
    def do_so         


        

Answer 1

I'm not sure, if this is a solution. You can instantiate objects in the apply method. I created an example class and dataframe.

class Test:
    def __init__(self, a, b):
        self.a = a
        self.b = b

df = pd.DataFrame({
    'group': list('abbaabcc'),
    'group2': list('abababab'),
    'a': [1,2,1,2,3,2,3,4],
    'b': [3,4,2,3,4,5,3,4]
})
df

Output

  group group2  a  b
0     a      a  1  3
1     b      b  2  4
2     b      a  1  2
3     a      b  2  3
4     a      a  3  4
5     b      b  2  5
6     c      a  3  3
7     c      b  4  4

Create Objects in apply

df.groupby(['group','group2'])[['a','b']].apply(
    lambda x: [Test(e[0],e[1]) for _,e in x.iterrows()])

Output

group  group2
a      a         [<__main__.Test object at 0x7f5351df0390>, <__...
       b                [<__main__.Test object at 0x7f5351df03d0>]
b      a                [<__main__.Test object at 0x7f5351df0450>]
       b         [<__main__.Test object at 0x7f5351df0490>, <__...
c      a                [<__main__.Test object at 0x7f5351df0550>]
       b                [<__main__.Test object at 0x7f5351df04d0>]
dtype: object