Small traps in Perl functions that return lists
I often find some code that works as follows: given a list of items, there is a function that provides a list of unique items, and another function that counts how many unique items are there. The code looks like the following:#!perl
use v5.38;
sub unique_list {
my $unique = {};
$unique->{ $_ }++ for ( @_ );
return sort keys $unique->%*;
}
sub unique_count {
return scalar( unique_list( @_ ) ) // 0;
}
The problem is within
unique_count: the function does not “store” the result of unique_list into a list, but passes it to scalar, that results in a zero or undef value. Just think at what happens if the unique_list returns a single value.
Assuming the following code:
my @animals = qw/ cat dog fish cat dog bird /;
say unique_list @animals;
say unique_count( @animals );
The
unique_count will always return zero, even if unique_list is working as expected.
It is Perl!
There are several solutions to this, one could be to assign the result of
unique_list to a list:
sub unique_count {
my @unique = unique_list( @_ );
scalar( @unique );
}
or use an array reference when dealing with
unique_list:
sub unique_list {
my $unique = {};
$unique->{ $_ }++ for ( @_ );
return [ sort keys $unique->%* ];
}
sub unique_count {
return scalar( unique_list( @_ )->@* ) // 0;
}
Clearly, the latter solution means that users of
unique_list has to de-reference the returned value before it can be used:
say unique_list( @animals )->@*;
say unique_count( @animals );
The solution to adopt is pretty much a matter of taste, I tend to prefer the array ref solution because it makes much more clear that I have to derefence an array, so there are less chance I’m going to forget to use a list in the counting-zone.