The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.
NAME
    Data::Dmap - just like map, but on deep data structures

VERSION
    Version 0.07.

SYNOPSIS
    This module provides the single function "dmap" which carries out a
    "map"-like operation on deep data structures.

        use Data::Dmap;

        my $foo = {
            cars => [ 'ford', 'opel', 'BMW' ],
            birds => [ 'cuckatoo', 'ostrich', 'frigate' ],
            handler => sub { print "barf\n" }
        };

        # This removes all keys named 'cars'    
        my($bar) = dmap { delete $_->{cars} if ref eq 'HASH'; $_ } $foo;

        # This replaces arrays with the number of elements they contains
        my($other) = dmap { $_ = scalar @$_ if ref eq 'ARRAY'; $_ } $foo;

        use Data::Dumper;
        print Dumper $other;
        #
        # Prints
        # {
        #    birds => 3,
        #    handler => sub { "DUMMY" }
        # }
        # (Data::Dumper doesn't dump subs)

        $other->{handler}->();
        # Prints
        # barf

EXPORTS
    "dmap" (always exported) - the dmap function that does deep in-place
    mapping
    "cut" (optional) - a function for stopping recursion.

SUBROUTINES
  "dmap"
    This function works like "map" - it takes an expression followed by a
    list, evaluates the expression on each member of the list and returns
    the result.

    The only difference is that any references returned by the expression
    will also be traversed and passed to the expression once again, thus
    making it possible to make deep traversal of any data structure.

    Objects (references blessed to something) are just traversed as if they
    weren't blessed.

   Examples
    Delete all hash references

        use Data::Dmap;
        use Data::Dump 'pp';
        
    pp dmap { return $_ unless ref eq 'HASH'; return; } 1, 'foo', [ { a => 1 }, 2];
        
    # Prints:
        # (1, "foo", [2])

    Delete every odd number

        use Data::Dmap;
        use Data::Dump 'pp';
        
    pp dmap { return if $_ % 2; $_ } [ 1 .. 10 ];

        # Prints:
        # [2, 4, 6, 8, 10]

    Replace all hash refs with some $object of class "thingy".

        use Data::Dmap;
        use Data::Dump 'pp';
        
    pp dmap { return bless $_, 'thingy' if ref eq 'HASH'; $_ } [ 1, "hello", { a => 1 } ];
        
    # Prints:
        # [1, "hello", bless({ a => 1 }, "thingy")]

    "dmap" understands what you want, if you return nothing (as opposed to
    "undef") when evaluating the expression for a hash key:

        use Data::Dmap;
        use Data::Dump 'pp;

        my $characters = { main => 'pooh', secondary => 'piglet' };
        pp dmap { return if $_ eq "piglet"; $_ } $characters;
        
    # Prints:
        # { main => "pooh" }

    Because the output from the expression is being traversed, you can use
    "dmap" to generate data structures:

        use Data::Dmap;
        use Data::Dump 'pp';
        
    my $height = 3;
        pp dmap { if(ref eq 'HASH' and $height--) { $_->{a} = {height => $height} } $_ } {};
        
    # Prints:
        # {
        #     a => {
        #         a => {
        #             a => { 
        #                 height => 0
        #             },
        #             height => 1
        #         },
        #         height => 2
        #     }
        # }
        # (My own formatting above.)

  "cut"
    The "cut" routine stops recursion at any point and returns any data as
    it is in place of the current node.

   Examples
        use Data::Dmap 'cut';
        use Data::Dump 'pp';
        
    my $deep = {
            level => 1,
            data  => {
                level => 2,
                data => {
                    level => 3
                }
            }
        };
        
    pp dmap { cut('stop') if ref eq 'HASH' and $_->{level} == 2} $deep;

        # Prints:
        #
        # { data => { data => "stop", level => 2 }, level => 1 }

AUTHOR
    Michael Zedeler, "<michael@zedeler.dk>"

BUGS
    If you find a bug, please consider helping to fix the bug by doing this:

    *   Fork "Data::Dmap" from <http://github.com/mzedeler/Data-Dmap>

    *   Write a test case in the "t" directory, commit and push it.

    *   Fix the bug or (if you don't know how to fix it), report the bug

    Bugs and feature requests can be reported through the web interface at
    <http://github.com/mzedeler/Data-Dmap/issues>. I may not be notified, so
    send me a mail too.

SUPPORT
    You can find documentation for this module with the perldoc command.

        perldoc Data::Dmap

    You can also look for information at:

    *   The github issue tracker

        <http://github.com/mzedeler/Data-Dmap/issues>

    *   AnnoCPAN: Annotated CPAN documentation

        <http://annocpan.org/dist/Data-Dmap>

    *   CPAN Ratings

        <http://cpanratings.perl.org/d/Data-Dmap>

    *   Search CPAN

        <http://search.cpan.org/dist/Data-Dmap/>

SEE ALSO
    Data::Rmap, Data::Visitor, Data::Transformer, Data::Visitor, Data::Walk.

TODO
    Some kind of option making it possible to traverse objects with
    Class::MOP metaclasses, so we can avoid breaking encapsulation.
    Options to provide more information about the current node to the
    callback handler, such as path, depth and data types. Should not affect
    performance if not used.

LICENSE AND COPYRIGHT
    Copyright 2010 Michael Zedeler.

    This program is free software; you can redistribute it and/or modify it
    under the terms of either: the GNU General Public License as published
    by the Free Software Foundation; or the Artistic License.

    See http://dev.perl.org/licenses/ for more information.