That way is fine. I might have mapped instead:

my %rounded_stats = map {
    $_ => sprintf '%.2f', $stats{$_}
} keys %stats;
[download]

Another way to do the actual rounding is to use Math::Round. The nearest() function can round to a whatever acuracy you want. This does the same as the sprintf() version, except that the results are numbers, so you don't get trailing '0' characters in the output:

my %rounded_stats = map {
    $_ => nearest(0.01, $stats{$_})
} keys %stats;
[download]

It seemed long. I need to think about using map more. thanks

$_= sprintf "%.2f", $_
    for values( my %rounded= %stats );
[download]

But that won't work in old Perls where values doesn't return a list of aliases.

While I must second the use of map and sprintf, you may be interested to know that rounding constitutes several Categorized Questions and Answers, is found in perlfaq4, and a fairly common question. In the interest of completness (and TimToady), there are other methods to round numbers as shown in a variety of threads (eg Re: Rounding function and Re^2: Splitting String???), which include the following:

int ($num + .5);

($num += 5 / (10 ** ($decimal_places + 1))) =~ s/(?<=\.)(\d{1,$decimal
+_places}).*$/$1/;
[download]

In addition to Zaxo's use of map, you may also be interested in using $, (or $" if interpolated; see perlvar), and printing directly, eg:

$, = $/;
print map { sprintf("%.2f", $_) } values %stats;
[download]

Here's another idea: use tie.

To explain myself some more:

Create a module, in there you create a hash, and make the tie method accept a param called 'rounded', which has a true/false value.

Then tie it to two hashes, in your example, you tie the hash %stats to the module, with the rounded param set to false, and tie the hash %rounded_stats to the same module, but with the rounded param set to true.

Then implement the FETCH, and VALUES method to check the value of the rounded-param...

This would have the advantage that your data is only stored once (this also means it is in sync at all time), but it would also mean that fetching a value might be a bit slower/less efficient (since it has to do the sprintf)...

I can't tell what would be best, creating a real hash vs tieing it to two hashes, since that really depends on how you are planning on using it...

(References: `perldoc -f tie`, `perldoc perltie`, and `perldoc Tie::Hash`)

Animator's text gave me inspiration to go create a tied hash module just for the learning experience.
Included below.

Read more... (2 kB)