current community

your communities

more stack exchange communities

Stack Exchange
tour help
stack overflow careers
Take the 2-minute tour ×
Stack Overflow is a question and answer site for professional and enthusiast programmers. It's 100% free, no registration required.
up vote 0 down vote favorite

are the both approaches equivalent from a point of view of performance and coding-style?

def foo(n):
    return n*2

# case 1 with ".append(function())"

mylist = [1,2,3,4,5,6,7,8,9,10]
result = list()
for l in mylist:
    result.append(foo(l))

# case 2 ".append(result)"

mylist = [1,2,3,4,5,6,7,8,9,10]
result = list()
for l in mylist:
    r = foo(l)
    result.append(r)
share|improve this question
1  
Solution 1 is more readable and compact but basically nobody cares. –  Andreas Jung Mar 10 '13 at 17:52
1  
performance = , coding != first better, but Python supports much better ways.. –  Grijesh Chauhan Mar 10 '13 at 17:52

4 Answers 4

up vote 3 down vote accepted

Your approaches are more or less equivalent; option 2 does a little more work; it has to do an extra store and name lookup for the temporary variable. In practice, the difference is minimal and you should be focusing on code readability more than premature optimization.

For your specific example you'd be much better off using a list comprehension:

result = [foo(l) for l in mylist]

The list comprehension constructs the list in C code and does not have to look up the .append() method each loop iteration.

share|improve this answer
    
Thanks Martijn. My question was more theoretical than practice. Normally i use the first case. –  Gianni Spear Mar 10 '13 at 17:54
    
Sorry for the extra question, but why C or C++ are more fast than other language as Python? –  Gianni Spear Mar 10 '13 at 18:00
    
@Gianni: Because C and C++ can be compiled to machine code, having made very different decisions about code structure and rigidity of the type system. In the end, what are you more productive in, what makes you create solutions faster? –  Martijn Pieters♦ Mar 10 '13 at 18:04
1  
@Gianni: In that light, this answer as to why Google uses Python is interesting. –  Martijn Pieters♦ Mar 10 '13 at 18:05
1  
Take a look at numpy, scipy, or pandas; all libraries that make specific analysis jobs much faster, while still giving you the power of Python to control it all. –  Martijn Pieters♦ Mar 10 '13 at 18:21
up vote 2 down vote

I think they are the same, but best option would be map(foo, mylist) - even from performance point of view as append can cause list resize which takes extra time

share|improve this answer
    
In python 3, map() would return an iterator; the list comprehension case can still outperform list(map(..)) if you needed a list all along. –  Martijn Pieters♦ Mar 10 '13 at 17:57
    
You are right with this point about Python 3. I think that performance difference between list(imap(..)) and map(..) would be very hard to notice especially for list longer than some (not big) X. Also for Python 2.x advantage could be on map side –  kkonrad Mar 10 '13 at 18:09
    
Yes, map() in Python 2 is slightly faster for the specific example inputs in the OP. With long lists, the list comprehension wins however, as it builds the result list more efficiently. –  Martijn Pieters♦ Mar 10 '13 at 18:14
    
Are you sure about that list is built differently for list comprehension than from some iterable? Intuitively this should work exactly the same way - that comprehension is transformed into generator which is base for new list... –  kkonrad Mar 10 '13 at 18:22
    
The list comprehension does not use a generator, actually. Both map() and a list comprehension have dedicated C code to build lists. I used the timeit module to compare the performance of both approaches. –  Martijn Pieters♦ Mar 10 '13 at 18:23
up vote 1 down vote 
result = [foo(i) for i in mylist]

is what you want.

share|improve this answer
up vote 2 down vote

You might have some performance gain in just removing the tempoarary assignment, but remember, premature optimization is evil, and micro-optimization is not worth the effort

Off course there are better ways to do, for example

List Comprehension

mylist = [1,2,3,4,5,6,7,8,9,10]
result = [n*2 for n in mylist]
share|improve this answer
    
Thanks Abhijit. Why "premature optimization is evil"? i read this sentence everywhere!!! –  Gianni Spear Mar 10 '13 at 18:44
1  
@Gianni You should be worrying about the clarity and robustness of the code first –  jamylak Mar 23 '13 at 14:27

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.