Sorting

An extremely useful builtin function in Perl is sort, which sorts an array. The default behaviour of sort is to use alphabetic sort. So, for instance, to read a list of phone numbers where each line is of the form name:phone number and print them out in alphabetical order we could write:

  while ($line = <PHONE>){
    ($name,$phone) = split /:/,$line;   
    $phonehash{$name} = $phone;       # Store phone number in a hash
  }

  foreach $k (sort keys %phonehash){
     print $k, ":", $phonehash{$k},"\n");
  }

Here, we sort the list generated by keys %phonehash before printout out the values in the hash.

What if we want to supply a different basis for comparing values? We can supply sort with the name of a comparison function. Then comparison function we supply will be invoked by sort. Instead of using @_ to pass parameters to the comparison function (as with normal functions), sort will pass the values as $a and $b. The comparison function should return $-1$ if the first argument, $a, is smaller than the second argument, $b$, 0 if the two arguments are equal, and 1 if the second argument is smaller than the first. So, we can sort using a numeric comparison function as follows.

   foreach $k (sort bynumber keys %somehash){
     ...
   }

   sub bynumber {
     if ($a < $b) {return -1};
     if ($a > $b) {return 1};
     return 0;
   }

In fact, this is so common that Perl supplies a builtin ``spaceship'' operator <=> that has this effect.

   sub bynumber {
     return $a <=> $b;  # Return -1 if $a < $b,
                        #        +1 if $a > $b,
                        #         0 if $a == $b
   }

The operator cmp achieves the same effect as <=> but using string comparison instead of numeric comparison.

To sort in descending order, we simply invert the position of the arguments in the comparison function.

   foreach $k (sort bynumdesc keys %somehash){
     ...
   }

   sub bynumberdesc {return $b <=> $a;}

We can also use the arguments $a and $b in more complex ways. For instance, to print out a hash based on the numeric sorted order of its values, we can write:

   foreach $k (sort bystrvalue keys %somehash){
     ...
   }

   sub bystrvalue {return $somehash{$a} cmp $somehash{$b};}

Finally, we can avoid using a separate named comparison function by just supplying the expression that is to be evaluated in braces, as follows:

   foreach $k (sort {$a <=> $b} keys %somehash){  # Same as bynumber
   foreach $k (sort {$b <=> $a} keys %somehash){  # Same as bynumdesc
   foreach $k (sort {$somehash{$a} cmp $somehash{$b}} 
                      keys %somehash){            # Same as bystrvalue