what's the difference between pandas DataFrame methods agg() and apply()?

Question

There are a number of SO questions regarding agg and apply on pandas DataFrame.groupby() objects, but I don't understand the difference between DataFrame.agg() and DataFrame.apply(). From the docs and the snippet below, they look the same to me. If there are issues specifically related to row operations that don't apply to operations on columns, I'd like to know about them.

import pandas as pd

a = pd.Series([True, False, False])
b = pd.Series([False, False, False])
c = pd.Series([True, True, False])
d = pd.Series([1, 2, 3])

print(pd.DataFrame({'a': a, 'b': b, 'c': c, 'd': d}).agg(lambda x: print(len(x)), axis=1))
print()
print(pd.DataFrame({'a': a, 'b': b, 'c': c, 'd': d}).apply(lambda x: print(len(x)), axis=1))

4
4
4
0    None
1    None
2    None
dtype: object

4
4
4
0    None
1    None
2    None
dtype: object

scotscotmcc · Accepted Answer · 2022-11-14 23:34:53Z

1

They both actually call the same frame_apply(...) function with generally the same things going into them. One difference is that df.agg() (which actually itself is identical to df.aggregate()) has a step where it calls a reconstruct_func() function that does a little bit of cleanup/handling of what func you pass in.

For example, you can do df['col1'].agg(sum) and get a result, but if you do df['col1'].apply(sum) you'll get an error. Both .agg() and .apply() will work with the string 'sum', though - so we see that .agg() can take a little more variety. (This is all with a column of integers)

answered Nov 14, 2022 at 23:34

scotscotmcc

3,1631 gold badge10 silver badges35 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

shortorian Over a year ago

but df['col1'] is a series not a dataframe, right?

scotscotmcc Over a year ago

oh, good catch... and now that I look at that, I can call df.agg(sum) or df.apply(sum) both without any problems. That is, apply on the df seems a little more robust than on the series itself, and makes my point there pretty moot... I mean, .agg() does call a reconstruct_func(), but doesn't seem to have the effect I said it did... I just spent some more time looking through the code and my conclusion is - not clear to me...

Collectives™ on Stack Overflow

what's the difference between pandas DataFrame methods agg() and apply()?

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related