Performance of python when passing big data as function parameter

Question

Suppose I have two Python functions, Function 1 and Function 2.

Function 1 will call Function 2 and the parameter is a big data (e.g., a dictionary with 100 thousand elements).

I am wondering if there is any performance differences between calling Function 2 in Function 1, which means I need to pass the big data parameter, and implementing Function 2 in Function 1 directly, which means I don't need to pass the big data parameter.

Thanks.

PS: I think the key question is how Python pass the parameter, by value, or by reference (pointer)?

Edit: It seems that this is a confused problem. How do I pass a variable by reference? is a good answer.

So what were the results when you tried both options and timed them, again? :) — Lev Levitsky, May 09 '13 at 22:31
@LevLevitsky Testing is fine but quite often a precise answer on inner workings is preferable — Basic, May 09 '13 at 22:37
@LevLevitsky This question come to my mind when I try to implement Function 2. I will implement both ways and see if there is any difference. thx — mitchelllc, May 09 '13 at 22:37

score 27 · Accepted Answer · edited May 23 '17 at 12:02

27

Python passes references-to-objects by value. The terminology is controversial and ugly, but there should be no real performance difference.

Check out these answers for all the details you could ever want (hopefully).

edited May 23 '17 at 12:02

Community

1
1

answered May 09 '13 at 22:35

Henry Keiter

16,863
7
51
80

2

(and all the details you hoped you never need to know) – gies0r Apr 10 '20 at 13:31

score 7 · Answer 2 · answered May 09 '13 at 22:36

7

The terminology of how python "passes" is a vicious debate I don't want to go into. But what is actually passed on the stack is a reference. So there is not a large memory cost with either of your options.

answered May 09 '13 at 22:36

cmd

5,754
16
30

Performance of python when passing big data as function parameter

2 Answers2