Data::Walk(3pm) User Contributed Perl Documentation Data::Walk(3pm)
NAME
Data::Walk - Traverse Perl data structures
SYNOPSIS
use Data::Walk;
walk &wanted, @items_to_walk;
use Data::Walk;
walkdepth &wanted, @items_to_walk;
use Data::Walk;
walk { wanted => &process, follow => 1 }, $self;
DESCRIPTION
The above synopsis bears an amazing similarity to File::Find(3pm) and this is not coincidental.
Data::Walk(3pm) is for data what File::Find(3pm) is for files. You can use it for rolling your own serialization class, for displaying
Perl data structures, for deep copying or comparing, for recursive deletion of data, or ...
If you are impatient and already familiar with File::Find(3pm), you can skip the following documentation and proceed with "DIFFERENCES TO
FILE::FIND".
FUNCTIONS
The module exports two functions by default:
walk
walk &wanted, @items;
walk \%options, @items;
As the name suggests, the function traverses the items in the order they are given. For every object visited, it calls the &wanted
subroutine. See "THE WANTED FUNCTION" for details.
walkdepth
walkdepth &wanted, @items;
walkdepth \%options, @items;
Works exactly like "walk()" but it first descends deeper into the structure, before visiting the nodes on the current level. If you
want to delete visited nodes, then "walkdepth()" is probably your friend.
OPTIONS
The first argument to "walk()" and "walkdepth()" is either a code reference to your &wanted function, or a hash reference describing the
operations to be performed for each visited node.
Here are the possible keys for the hash.
wanted
The value should be a code reference. This code reference is described in "THE WANTED FUNCTION" below.
bydepth
Visits nodes on the current level of recursion only after descending into subnotes. The entry point "walkdepth()" is a shortcut for
specifying "{ bydepth => 1 }".
preprocess
The value should be a code reference. This code reference is used to preprocess the current node $Data::Walk::container. Your prepro-
cessing function is called before the loop that calls the "wanted()" function. It is called with a list of member nodes and is
expected to return such a list. The list will contain all sub-nodes, regardless of the value of the option follow! The list is nor-
mally a shallow copy of the data contained in the original structure. You can therefore safely delete items in it, without affecting
the original data. You can use the option copy, if you want to change that behavior.
The behavior is identical for regular arrays and hashes, so you probably want to coerce the list passed as an argument into a hash
then. The variable $Data::Walk::type will contain the string "HASH" if the currently inspected node is a hash.
You can use the preprocessing function to sort the items contained or to filter out unwanted items. The order is also preserved for
hashes!
preprocess_hash
The value should be a code reference. The code is executed right after an eventual preprocess_hash handler, but only if the current
container is a hash. It is skipped for regular arrays.
You will usually prefer a preprocess_hash handler over a preprocess handler if you only want to sort hash keys.
postprocess
The value should be a code reference. It is invoked just before leaving the currently visited node. It is called in void context with
no arguments. The variable $Data::Walk::container points to the currently visited node.
follow
Causes cyclic references to be followed. Normally, the traversal will not descend into nodes that have already been visited. If you
set the option follow to a truth value, you can change this behavior. Unless you take additional measures, this will always imply an
infinite loop!
Please note that the &wanted function is also called for nodes that have already been visited! The effect of follow is to suppress
descending into subnodes.
copy
Normally, the &preprocess function is called with a shallow copy of the data. If you set the option copy to a false value, the &pre-
process function is called with one single argument, a reference to the original data structure. In that case, you also have to return
a suitable reference.
Using this option will result in a slight performance win, and can make it sometimes easier to manipulate the original data.
What is a shallow copy? Think of a list containing references to hashes:
my @list = ({ foo => 'bar' }, { foo => 'baz' });
my @shallow = @list;
After this, @shallow will contain a new list, but the items stored in it are exactly identical to the ones stored in the original. In
other words, @shallow occupies new memory, whereas both lists contain references to the same memory for the list members.
All other options are silently ignored.
THE WANTED FUNCTION
The &wanted function does whatever verifications you want on each item in the data structure. Note that despite its name, the &wanted
function is a generic callback and does not tell Data::Walk(3pm) if an item is "wanted" or not. In fact, its return value is ignored.
The wanted function takes no arguments but rather does its work through a collection of variables:
$_ The currently visited node. Think "file" in terms of File::Find(3pm)!
$Data::Walk::container
The node containing the currently visited node, either a reference to a hash or an array. Think "directory" in terms of
File::Find(3pm)!
$Data::Walk::type
The base type of the object that $Data::Walk::container references. This is either "ARRAY" or "HASH".
$Data::Walk::seen
For references, this will hold the number of times the currently visited node has been visited before. The value is consequently set
to 0 not 1 on the first visit. For non-references, the value is undefined.
$Data::Walk::address
For references, this will hold the memory address it points to. It can be used as a unique identifier for the current node. For non-
references, the value is undefined.
$Data::Walk::depth
The depth of the current recursion.
These variables should not be modified.
DIFFERENCES TO FILE
::FIND
The API of Data::Walk(3pm) tries to mimic the API of File::Find(3pm) to a certain extent. If you are already familiar with File::Find(3pm)
you will find it very easy to use Data::Walk(3pm). Even the documentation for Data::Walk(3pm) is in parts similar or identcal to that of
File::Find(3pm).
Analogies
The equivalent of directories in File::Find(3pm) are the container data types in Data::Walk(3pm). Container data types are arrays (aka
lists) and associative arrays (aka hashes). Files are equivalent to scalars. Wherever File::Find(3pm) passes lists of strings to func-
tions, Data::Walk(3pm) passes lists of variables.
Function Names
Instead of "find()" and "finddepth()", Data::Walk(3pm) uses "walk()" and "walkdepth()", like the smart reader has already guessed after
reading the "SYNOPSIS".
Variables
The variable $Data::Walk::container is vaguely equivalent to $File::Find::dir. All other variables are specific to the corresponding mod-
ule.
Wanted Function
Like its archetype from File::Find(3pm), the wanted function of Data::Walk(3pm) is called with $_ set to the currently inspected item.
Options
The option follow has the effect that Data::Walk(3pm) also descends into nodes it has already visited. Unless you take extra measures,
this will lead to an infinite loop!
A number of options are not applicable to data traversion and are ignored by Data::Walk(3pm). Examples are follow_fast, follow_skip,
no_chdir, untaint, untaint_pattern, and untaint_skip. To give truth the honor, all unrecognized options are skipped.
You may argue, that the options untaint and friends would be useful, too, allowing you to recursively untaint data structures. But, hey,
that is what Data::Walk(3pm) is all about. It makes it very easy for you to write that yourself.
EXAMPLES
Following are some recipies for common tasks.
Recursive Untainting
sub untaint {
s/(.*)/$1/s unless ref $_;
};
walk &untaint, $data;
See perlsec(1), if you don't understand why the untaint() function untaints your data here.
Recurse To Maximum Depth
If you want to stop the recursion at a certain level, do it as follows:
my $max_depth = 20;
sub not_too_deep {
if ($Data::Walk::depth > $max_depth) {
return ();
} else {
return @_;
}
}
sub do_something1 {
# Your code goes here.
}
walk { wanted => &do_something, preprocess => ¬_too_deep };
BUGS
If you think you have spotted a bug, you can share it with others in the bug tracking system at
http://rt.cpan.org/NoAuth/Bugs.html?Dist=Data-Walk.
COPYING
Copyright (C) 2005-2006, Guido Flohr <guido@imperia.net>, all rights reserved.
This program is free software; you can redistribute it and/or modify it under the terms of the GNU Library General Public License as pub-
lished by the Free Software Foundation; either version 2, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MER-
CHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Library General Public License for more details.
You should have received a copy of the GNU Library General Public License along with this program; if not, write to the Free Software Foun-
dation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
SEE ALSO
Data::Dumper(3pm), Storable(3pm), File::Find(3pm), perl(1)
perl v5.8.8 2007-09-17 Data::Walk(3pm)