Pandas DataFrame as an Argument to a Function - Python

Question

Suppose a Pandas DataFrame is passed to a function as an argument. Then, does Python implicitly copy that DataFrame or is the actual DataFrame being passed in?

Hence, if I perform operations on the DataFrame within the function, will I be changing the original (because the references are still intact)?

I just want to know whether or not I should make a Deep Copy of my DataFrame before passing it into a function and operating on it.

pretty good answer here: https://stackoverflow.com/questions/38895768/python-pandas-dataframe-is-it-pass-by-value-or-pass-by-reference — JohnE, Dec 03 '20 at 03:47

score 15 · Accepted Answer · edited Feb 12 '22 at 23:12

15

If a function parameter is a mutable object (e.g. a DataFrame), then any changes you make in the function will be applied to the object.

E.g.

In [200]: df = pd.DataFrame({1:[1,2,3]})

In [201]: df
Out[201]:
   1
0  1
1  2
2  3

In [202]: def f(frame):
     ...:     frame['new'] = 'a'
     ...:

In [203]: f(df)

In [204]: df
Out[204]:
   1 new
0  1   a
1  2   a
2  3   a

See this article for a good explanation on how Python passes function parameters.

edited Feb 12 '22 at 23:12

smci

29,564
18
109
144

answered Jul 17 '18 at 23:30

aydow

3,456
2
21
37

6

dataframes are *mutable*, not *immutable* – Paul H Apr 16 '19 at 15:46
@PaulH maybe op said DataFrame is an example of *not* immutable (=mutable) objects – David Johns Jan 09 '21 at 04:59
@ DavidJohns: the way that's written is very ambiguous e.g. *"If arg is not a bear (e.g. a a koala)*. Much clearer to say *If it's a mutable object..."* – smci Feb 12 '22 at 23:12

Pandas DataFrame as an Argument to a Function - Python

1 Answers1