我应该在Perl中使用哪个“for”循环？

Question

What are the arguments for using one of the for loops over the other to count through integers?

for my $i (0..$n) { ... }

for (my $i = 0; $i <= $n; $i++) { ... }

Answer 1

In your case, there isn't any difference.

In other use cases, in a for-each loop you don't have control of or access to the iterator.

Using your example syntax:

for my $i (0..10)
{
   print "$i";
   $i++;
}

The above actually creates a foreach loop - it's the same as saying foreach my $i (0..10) . $i is a value returned from the list , not the iterator. The iterator is internal and you don't have any access to it; you can not control the flow of the loop.

The output from above will print 012345678910 .

This:

for ( my $i = 0; $i++ ; $i <= 10)
{
    print $i;
    $i++;
}

That is an actual for loop. You are controlling and outputting the iterator. It will output: 0246810

In addition:

When you do for (1..$n) you are invoking the range operator vs. doing a simple comparison at the top of the loop. Performance differences, however, would be immeasurable.

Answer 2

The difference between the two styles of for loops in Perl is one of both clarity and efficiency.

When you look at for my $i (0 .. $n) {...} you can instantly see the range being used without having to mentally parse a larger expression.

With for (my $i = 0; $i <= $n; $i++) {...} there is quite a bit more to look at, and more places where errors can creep in.

In addition, foreach over a range is faster than the equivalent C-style loop as shown by the following benchmark:

use Benchmark 'cmpthese';

for my $mag (map 10**$_, 1 .. 6) {
    print "\n$mag:\n";
    cmpthese -2 => {
        loop => sub {my $x = 0; for (my $i = 0; $i <= $mag; $i++) {$x += $i}},
        each => sub {my $x = 0; for my $i (0 .. $mag) {$x += $i}},
    };
}

which prints:

10:
         Rate loop each
loop 613877/s   --  -2%
each 625568/s   2%   --

100:
         Rate loop each
loop  79481/s   -- -24%
each 104758/s  32%   --

1000:
        Rate loop each
loop  8140/s   -- -27%
each 11220/s  38%   --

10000:
       Rate loop each
loop  832/s   -- -26%
each 1124/s  35%   --

100000:
       Rate loop each
loop 81.6/s   -- -26%
each  110/s  34%   --

1000000:
       Rate loop each
loop 6.90/s   -- -26%
each 9.27/s  34%   --

Answer 3

With the three-part for loop, you are able to $i += 3 and such if you want from within the loop, but with the foreach loop, you are not. That is the difference.

Answer 4

The off-by-one error that you originally had in the C-style loop shows that the C-style loop is more error prone, and should only be used if you actually need access to the loop counter. A foreach loop makes it much more difficult to make an off-by-one error.

Answer 5

I believe that the two loop examples you posted behave identically. In both cases, $i is lexically scoped (since you use the my keyword). I think the 0..$n syntax is clearer, but (surprise!) not everyone would agree with that.