How can I use `apply` with a function that takes multiple inputs

Question

I have a function that has multiple inputs, and would like to use SFrame.apply to create a new column. I can't find a way to pass two arguments into SFrame.apply.

Ideally, it would take the entry in the column as the first argument, and I would pass in a second argument. Intuitively something like...

def f(arg_1,arg_2):
    return arg_1 + arg_2

sf['new_col'] = sf.apply(f,arg_2)

score 21 · Answer 1 · edited Aug 28 '16 at 09:15

21

suppose the first argument of function f is one of the column.

Say argcolumn1 in sf, then

sf['new_col'] = sf['argcolumn1'].apply(lambda x:f(x,arg_2))

should work

edited Aug 28 '16 at 09:15

Ahsanul Haque

10,676
4
41
57

answered Oct 14 '15 at 15:01

Yuan Huang

311
1
4

please replace f with your function name like if you have function named count_imp_words then you will write: products['awesome'] = products['word_count'].apply(lambda x:count_imp_words(x,'awesome')) – Rajesh Goel Feb 04 '17 at 01:38

score 2 · Answer 2 · answered Oct 13 '15 at 17:56

2

Try this.

sf['new_col'] = sf.apply(lambda x : f(arg_1, arg_2))

answered Oct 13 '15 at 17:56

druk

553
6
16

score 0 · Answer 3 · answered Dec 11 '18 at 19:38

The way i understand your question (and because none of the previous answers are marked as accepted), it seems to me that you are trying to apply a transformation using two different columns of a single SFrame, so:

As specified in the online documentation, the function you pass to the SFrame.apply method will be called for every row in the SFrame.

So you should rewrite your function to receive a single argument representing the current row, as follow:

def f(row):
    return row['column_1'] + row['column_2']

sf['new_col'] = sf.apply(f)

How can I use `apply` with a function that takes multiple inputs

3 Answers3