4

I would like to apply a function with argument to a pandas series: I have found two different solution of SO:

python pandas: apply a function with arguments to a series

and

Passing multiple arguments to apply (Python)

both of them rely on the use of functool.partial and they works absolutely fine. By the way the new version of Pandas support multiple argument: in any case I do not understand how does it works. Example:

a=pd.DataFrame({'x':[1,2],'y':[10,20]})
a['x'].apply(lambda x,y: x+y, args=(100))

It exits with a:

TypeError: <lambda>() argument after * must be a sequence, not int
Community
  • 1
  • 1
user2988577
  • 3,997
  • 7
  • 21
  • 21

1 Answers1

9

The TypeError is saying that you passed the wrong type to the lambda function x + y. It's expecting the args to be a sequence, but it got an int. You may have thought that (100) was a tuple (a sequence), but in python it's the comma that makes a tuple:

In [10]: type((100))
Out[10]: int

In [11]: type((100,))
Out[11]: tuple

So change your last line to

In [12]: a['x'].apply(lambda x, y: x + y, args=(100,))
Out[12]: 
0    101
1    102
Name: x, dtype: int64
TomAugspurger
  • 28,234
  • 8
  • 86
  • 69