Apply function to columns of pandas data frame

Question

This works correctly:

import pandas as pd

def fnc(m):
    return m+4

df = pd.DataFrame({"m": [1,2,3,4,5,6], "c": [1,1,1,1,1,1], "x":[5,3,6,2,6,1]})
df
# apply a self created function to a single column in pandas
df["y"] = df['m'].apply(fnc)
df

I tried to modified code above. Here I need to add column m value to column c value and assign result to column y:

import pandas as pd

def fnc(m,c):
    return m+c

df = pd.DataFrame({"m": [1,2,3,4,5,6], "c": [1,1,1,1,1,1], "x":[5,3,6,2,6,1]})
df
# apply a self created function to a single column in pandas
df["y"] = df[['m','c']].apply(fnc)
df

gives me error:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
d:\del\asfasf.py in 
      8 df
      9 # apply a self created function to a single column in pandas
----> 10 df["y"] = df[['m','c']].apply(fnc)
     11 df

C:\ProgramData\Anaconda3\lib\site-packages\pandas\core\frame.py in apply(self, func, axis, raw, result_type, args, **kwds)
   6876             kwds=kwds,
   6877         )
-> 6878         return op.get_result()
   6879 
   6880     def applymap(self, func) -> "DataFrame":

C:\ProgramData\Anaconda3\lib\site-packages\pandas\core\apply.py in get_result(self)
    184             return self.apply_raw()
    185 
--> 186         return self.apply_standard()
    187 
    188     def apply_empty_result(self):

C:\ProgramData\Anaconda3\lib\site-packages\pandas\core\apply.py in apply_standard(self)
    294             try:
    295                 result = libreduction.compute_reduction(
--> 296                     values, self.f, axis=self.axis, dummy=dummy, labels=labels
    297                 )
    298             except ValueError as err:

pandas\_libs\reduction.pyx in pandas._libs.reduction.compute_reduction()

pandas\_libs\reduction.pyx in pandas._libs.reduction.Reducer.get_result()

TypeError: fnc() missing 1 required positional argument: 'c'

Question: How to correct my second code? If possible, please provide answer standard function syntax (not lambda function)

score 4 · Accepted Answer · answered Jul 04 '20 at 04:00

4

Add axis to consider in axis=1 in dataframe and access each column inside the function.Try this

def fnc(m):
    return (m.m+m.c)

df = pd.DataFrame({"m": [1,2,3,4,5,6], "c": [1,1,1,1,1,1], "x":[5,3,6,2,6,1]})
df["y"] = df[['m',"c"]].apply(fnc,axis=1)

or you can apply to df,without select "m" and "c" columns.

df["y"] = df.apply(fnc,axis=1)

output

answered Jul 04 '20 at 04:00

Rajith Thennakoon

3,975
2
14
24

If I run `df` after your code I am receiving `y` column field with `None`. It should contain column with calculated values. – vasili111 Jul 04 '20 at 04:08
because in func it was print statement(mistake from me),and i changed it to return.now it should work. – Rajith Thennakoon Jul 04 '20 at 04:09
Thank you. Can you please add explanation of this `m.m+m.c` part? What it means? – vasili111 Jul 04 '20 at 04:14
1

its extract the m and c column values and add together row wise.since we defined axis=1 in the apply function it enumerates row wise – Rajith Thennakoon Jul 04 '20 at 04:19

score 2 · Answer 2 · answered Jul 04 '20 at 03:57

2

try this, set axis to 1 for column wise operation & *x for un-packing arguments.

df["y"] = df[['m','c']].apply(lambda x : fnc(*x), axis=1)

answered Jul 04 '20 at 03:57

sushanth

8,275
3
17
28

What you think, is there any advantage over your version and `df["y"] = df.apply(lambda row: fnc(row['m'], row['c']), axis = 1)`? – vasili111 Jul 04 '20 at 04:39
1

No, both are the same ``*x`` is implicit where as ``row['m'], row['c']`` is explicit. – sushanth Jul 04 '20 at 04:40
In your code, what is passed to `x` parameter of lambda function? I really like it and want to understand it. – vasili111 Jul 04 '20 at 05:39
Refer this, https://stackoverflow.com/a/890188/4985099 – sushanth Jul 04 '20 at 05:43

Apply function to columns of pandas data frame

2 Answers2