CODEWITHSUNDEEP
python
Shuwn Yuan Tee
Shuwn Yuan Tee

Reputation: 5748

Is there any reason using pass in favourable over None as empty statement

I came across a legacy code base, which is using pass, and None interchange, as empty statement.

if True:
    pass

if True:
    None

Since we prefer to have a consistent style over our code base, we wish to stick with one common standard. We need to choose between pass and None.

Python community mostly recommend using pass as empty statement.

https://stackoverflow.com/a/15081801/1051441

But, is there any side effect if we use None?

Is there any reason using pass in favourable over None as empty statement?

What is the difference between Pass and None in Python talks about difference among the 2. But, it doesn't mention why pass is preferable over None as empty statement.

Upvotes: 1

Views: 98

Answers (3)

Cong Ma
Cong Ma

Reputation: 11302

As other answers mentioned, there is no difference in the meaning of the code. Furthermore, recent versions of CPython enables dead code elimination on these kind of examples, and the bytecode is exactly the same in the OP's examples.

However, here's a very slight performance difference, in particular in earlier versions without optimization. Namely, when Python hits the pass keyword statement, it just jumps over it. But when it hits the None expression, it loads the value and then discards it, with the implied increment and decrement of reference count. Therefore, the latter should be slightly slower.

Example:

$ python -m timeit -r 10 -n 10000000 -s 'def f(): None' 'f()' 
10000000 loops, best of 10: 0.209 usec per loop
$ python -m timeit -r 10 -n 10000000 -s 'def f(): pass' 'f()' 
10000000 loops, best of 10: 0.178 usec per loop

Upvotes: 1

Eric
Eric

Reputation: 6066

Using pass is better.

Technically, in your example, there is no difference. Both snippets have the same disassembly:

(note that in fact, the function does absolutely nothing. The LOAD CONST None is for an implicit return statement, returning None.)

>>> def f():
...     if True:
...         pass
...
>>> def g():
...     if True:
...         None
...
>>> dis.dis(f)
  3           0 LOAD_CONST               0 (None)
              2 RETURN_VALUE
>>> dis.dis(g)
  3           0 LOAD_CONST               0 (None)
              2 RETURN_VALUE

However, it's important to consider readability of your code. The statement None and pass essentially mean different things. pass is a placeholder, None is a data value. It is merely a coincidence that having None by itself does nothing. It is better to code what you mean.

Upvotes: 2

Giacomo Catenazzi
Giacomo Catenazzi

Reputation: 9523

pass is explicit about what to you want. Python philosophy: Explicit is better than implicit., see the Zen of Python

Some IDEs warn you if you write some code after pass, but not after None, so you have an extra check. I don't know why python code doesn't warn about so obvious user error (I doesn't think user will use pass and write other instructions).

Upvotes: 0

Related Questions