How to compose a dataframe with return columns from result of applying lambda using two parameters of dataframe in Pandas

Question

I have a dataframe with 3 columns a,b and c, and a function that takes 3 parameters, for instance a small example:

data_test = [[1,11,101],[2,12,102],[3,13,103],[4,14,104],[5,15,105],[6,16,106]]
df_test = pd.DataFrame(data_test,columns=['a','b','c'],dtype=float)


      a        b        c
0   1.0     11.0    101.0
1   2.0     12.0    102.0
2   3.0     13.0    103.0
3   4.0     14.0    104.0
4   5.0     15.0    105.0
5   6.0     16.0    106.0


def my_function(a,b,c):
    #changes a b and c and returns 
    x = a*10
    y = b-20
    z = a*b -7
    return [x,y,z]

for each row I want to apply the function and return the values a,b,c,x,y,z in a new dataframe

I did:

df_wanted = pd.DataFrame( df_test.apply(lambda row: my_function(row['a'], row['b'], row['c']), axis=1) )

it is returning:

    0
0   [10.0, -9.0, 4.0]
1   [20.0, -8.0, 17.0]
2   [30.0, -7.0, 32.0]
3   [40.0, -6.0, 49.0]
4   [50.0, -5.0, 68.0]
5   [60.0, -4.0, 89.0]

How to get result like below instead of arrays for each row:

      a        b        c          x    y    z
0   1.0     11.0    101.0       10.0 -9.0  4.0
1   2.0     12.0    102.0       20.0 -8.0 17.0
2   3.0     13.0    103.0       30.0 -7.0 32.0
3   4.0     14.0    104.0       40.0 -6.0 49.0
4   5.0     15.0    105.0       50.0 -5.0 68.0
5   6.0     16.0    106.0       60.0 -4.0 89.0

Answer 1

You can return a panda series instead of an array:

def my_function2(a,b,c):
    #changes a b and c and returns 
    x = a*10
    y = b-20
    z = a*b -7
    return pd.Series({
        'x': x,
        'y': y,
        'z': z
    })

df_wanted = pd.concat([
    df_test,
    df_test.apply(lambda row: my_function2(row['a'], row['b'], row['c']), axis=1)
], axis=1)

I know the function in your example is likely trivial, but try using vectorized functions to operate on columns instead of row-by-row. It's a lot more efficient.

Answer 2

Fix your code

df=df_test.join( pd.DataFrame( df_test.apply(lambda row: my_function(row['a'], row['b'], row['c']), axis=1).tolist() ,columns=list('xyz')))