home | O'Reilly's CD bookshelfs | FreeBSD | Linux | Cisco | Cisco Exam  


Learning Perl Objects, References & ModulesLearning Perl Objects, References & ModulesSearch this book

5.5. Using map

The map operator has a very similar syntax to the grep operator and shares a lot of the same operational steps. For example, items from a list of values are temporarily placed into $_ one at a time, and the syntax allows both a simple expression form and a more complex block form.

However, the testing expression becomes a mapping expression. This expression is evaluated in a list context (not a scalar context). Each evaluation of the expression gives a portion of the many results. The overall result is the list concatenation of all individual results. (In a scalar context, map returns the number of elements that are returned in a list context. But map should rarely, if ever, be used in anything but a list context.)

Let's start with a simple example:

my @input_numbers = (1, 2, 4, 8, 16, 32, 64);
my @result = map $_ + 100, @input_numbers;

For each of the seven items placed into $_, you get a single output result: the number that is 100 greater than the input number, so the value of @result is 101, 102, 104, 108, 116, 132, and 164.

But you're not limited to having only one output for each input. Let's see what happens when each input produces two output items:

my @result = map { $_, 3 * $_ } @input_numbers;

Now there are two items for each input item: 1, 3, 2, 6, 4, 12, 8, 24, 16, 48, 32, 96, 64, and 192. Those pairs can be stored in a hash, if you need a hash showing what number is three times a small power of two:

my %hash = @result;

Or, without using the intermediate array from the map:

my %hash = map { $_, 3 * $_ } @input_numbers;

You can see that map is pretty versatile; you can produce any number of output items for each input item. And you don't always need to produce the same number of output items. Let's see what happens when you break apart the digits:

my @result = map { split //, $_ } @input_numbers;

Each number is split into its digits. For 1, 2, 4, and 8, you get a single result. For 16, 32, and 64, you get two results per number. When the lists are concatenated, you end up with 1, 2, 4, 8, 1, 6, 3, 2, 6, and 4.

If a particular invocation results in an empty list, that empty result is concatenated into the larger list, contributing nothing to the list. You can use this feature to select and reject items. For example, suppose you want only the split digits of numbers ending in 4:

my @result = map {
  my @digits = split //, $_;
  if ($digits[-1] == 4) {
    @digits;
  } else {
    ( );
  }
} @input_numbers;

If the last digit is 4, you return the digits themselves by evaluating @digits (which is evaluated in a list context). If the last digit is not 4, you return an empty list, effectively removing results for that particular item. (Thus, a map can always be used in place of a grep, but not vice versa.)

Of course, everything you can do with map and grep, you can also do with explicit foreach loops. But then again, you can also code in assembler or by toggling bits into a front panel.[24] The point is that proper application of grep and map can help reduce the complexity of the program, allowing you to concentrate on high-level issues rather than details.

[24]If you're old enough to remember those front panels.



Library Navigation Links

Copyright © 2003 O'Reilly & Associates. All rights reserved.