Re^2: calculate average from range of hash values (optimize)

Nor do you need to make a copy of the hash.

use List::Util qw(sum);
sub average {
    my ($data, $first, $last) = @_;
    my @goodkeys = grep exists $data->{$_}, $first .. $last;
    sum( @{$data}{@goodkeys} ) / @goodkeys;
}
[download]

Finally, that we're doing math on the values without checking their definedness suggests that valid values are never undef, in which case you can skip the double lookup.

use List::Util qw(sum);
sub average {
    my ($data, $first, $last) = @_;
    my @values = grep defined, @{$data}{$first .. $last};
    sum(@values) / @values;
}
[download]

That should be about as fast as you can get in Perl.

Makeshifts last the longest.

Comment on Re^2: calculate average from range of hash values (optimize) Select or Download Code

Replies are listed 'Best First'.
Re: Re^2: calculate average from range of hash values (optimize) by demerphq (Chancellor) on Jul 27, 2003 at 13:59 UTC
For the record your second solution corrupts the hash with keys whose value is undef. While the OP didnt explicitly say this was bad, I believe it was implicit in the comment that the keys could be very large or very small. Thus a search could inadverdantly create millions of keys and overflow available RAM. --- demerphq _{<Elian> And I do take a kind of perverse pleasure in having an OO assembly language...}	[reply] [d/l]
Re^4: calculate average from range of hash values (optimize) by Aristotle (Chancellor) on Jul 27, 2003 at 18:32 UTC
For the record your second solution corrupts the hash with keys whose value is undef. Eh? Extraordinarily strange. So a slice lookup autovivifies elements, while a single-element lookup doesn't. Feels like a bug, if you ask me.. certainly counterintuitive. Easy enough to fix, though. `use List::Util qw(sum); sub average { my ($data, $first, $last) = @_; my @values = map { local $_ = $data->{$_}; defined() ? $_ : () } $ +first .. $last; sum(@values) / @values; }` [download] Makeshifts last the longest.	[reply] [d/l]