These three expressions seem to be equivalent:
a,b,c = line.split()
(a,b,c) = line.split()
[a,b,c] = line.split()
Do they compile to the same code?
Which one is more pythonic?
1 Answer
According to dis, they all get compiled to the same bytecode:
>>> def f1(line):
... a,b,c = line.split()
...
>>> def f2(line):
... (a,b,c) = line.split()
...
>>> def f3(line):
... [a,b,c] = line.split()
...
>>> import dis
>>> dis.dis(f1)
2 0 LOAD_FAST 0 (line)
3 LOAD_ATTR 0 (split)
6 CALL_FUNCTION 0
9 UNPACK_SEQUENCE 3
12 STORE_FAST 1 (a)
15 STORE_FAST 2 (b)
18 STORE_FAST 3 (c)
21 LOAD_CONST 0 (None)
24 RETURN_VALUE
>>> dis.dis(f2)
2 0 LOAD_FAST 0 (line)
3 LOAD_ATTR 0 (split)
6 CALL_FUNCTION 0
9 UNPACK_SEQUENCE 3
12 STORE_FAST 1 (a)
15 STORE_FAST 2 (b)
18 STORE_FAST 3 (c)
21 LOAD_CONST 0 (None)
24 RETURN_VALUE
>>> dis.dis(f3)
2 0 LOAD_FAST 0 (line)
3 LOAD_ATTR 0 (split)
6 CALL_FUNCTION 0
9 UNPACK_SEQUENCE 3
12 STORE_FAST 1 (a)
15 STORE_FAST 2 (b)
18 STORE_FAST 3 (c)
21 LOAD_CONST 0 (None)
24 RETURN_VALUE
So they should all have the same efficiency. As far as which is most Pythonic, it's somewhat down to opinion, but I would favor either the first or (to a lesser degree) the second option. Using the square brackets is confusing because it looks like you're creating a list (though it turns out you're not).
2 Comments
Casey Falk
"Which one is more pythonic?" -- I'd say the first option since it has less syntax and is more "readable."
dano
@Kevin It's the same on Python 3.2, at least (that's all I have access to right now).
[[a]] = b, e.g., and((a,),) = bdoesn’t look particularly nice either. As for which one’s more readable when that’s not necessary, it’s up to you.