A simple hash is close to an array. Their initializations even look similar. First the array:
@last_name = (
"Ward", "Cleaver",
"Fred", "Flintstone",
"Archie", "Bunker"
);
Now let s represent the same information with a hash (aka associative array):
%last_name = (
"Ward", "Cleaver",
"Fred", "Flintstone",
"Archie", "Bunker"
);
Although they have the same name, the array @last_name
and the hash %last_name
are completely independent.
With the array, if we want to know Archie s last name, we have to perform a linear search:
my $lname;
for (my $i = 0; $i < @last_name; $i += 2) {
$lname = $last_name[$i+1] if $last_name[$i] eq "Archie";
}
print "Archie $lname
";
With the hash, it s much more direct syntactically:
print "Archie $last_name{Archie}
";
Say we want to represent information with only slightly richer structure:
- Cleaver (last name)
- Ward (first name)
- June (spouse s first name)
- Flintstone
- Bunker
Before references came along, flat key-value hashes were about the best we could do, but references allow
my %personal_info = (
"Cleaver", {
"FIRST", "Ward",
"SPOUSE", "June",
},
"Flintstone", {
"FIRST", "Fred",
"SPOUSE", "Wilma",
},
"Bunker", {
"FIRST", "Archie",
"SPOUSE", "Edith",
},
);
Internally, the keys and values of %personal_info
are all scalars, but the values are a special kind of scalar: hash references, created with {}
. The references allow us to simulate "multi-dimensional" hashes. For example, we can get to Wilma via
$personal_info{Flintstone}->{SPOUSE}
Note that Perl allows us to omit arrows between subscripts, so the above is equivalent to
$personal_info{Flintstone}{SPOUSE}
That s a lot of typing if you want to know more about Fred, so you might grab a reference as sort of a cursor:
$fred = $personal_info{Flintstone};
print "Fred s wife is $fred->{SPOUSE}
";
Because $fred
in the snippet above is a hashref, the arrow is necessary. If you leave it out but wisely enabled use strict
to help you catch these sorts of errors, the compiler will complain:
Global symbol "%fred" requires explicit package name at ...
Perl references are similar to pointers in C and C++, but they can never be null. Pointers in C and C++ require dereferencing and so do references in Perl.
C and C++ function parameters have pass-by-value semantics: they re just copies, so modifications don t get back to the caller. If you want to see the changes, you have to pass a pointer. You can get this effect with references in Perl:
sub add_barney {
my($personal_info) = @_;
$personal_info->{Rubble} = {
FIRST => "Barney",
SPOUSE => "Betty",
};
}
add_barney \%personal_info;
Without the backslash, add_barney
would have gotten a copy that s thrown away as soon as the sub returns.
Note also the use of the "fat comma" (=>
) above. It autoquotes the string on its left and makes hash initializations less syntactically noisy.