NAME
XML::Merge - flexibly merge XML documents
VERSION
This documentation refers to version 1.4 of XML::Merge, which was
released on Sat Jul 23 14:39:59:48 -0500 2016.
SYNOPSIS
#!/usr/bin/perl
use strict;use warnings;
use utf8;use XML::Merge;
# create new XML::Merge object from MainFile.xml
my $merge_obj= XML::Merge->new('filename' => 'MainFile.xml');
# Merge File2Add.xml into MainFile.xml
$merge_obj->merge( 'filename' => 'File2Add.xml');
# Tidy up the indenting that resulted from the merge
$merge_obj->tidy();
# Write out changes back to MainFile.xml
$merge_obj->write();
DESCRIPTION
This module inherits from XML::Tidy which in turn inherits from
XML::XPath. This ensures that Merge objects' indenting can be tidied up
after any merge operation since such modification usually ruins
indentation. Polymorphism allows Merge objects to be utilized as normal
XML::XPath objects as well.
The merging behavior is setup to combine separate XML documents
according to certain rules and configurable options. If both documents
have root nodes which are elements of the same name, the documents are
merged directly. Otherwise, one is merged as a child of the other. An
optional XPath location can be specified as the place to perform the
merge. If no location is specified, the merge is attempted at the first
matching element or is appended as the new last child of the other root
if no match is found.
USAGE
new()
This is the standard Merge object constructor. It can take the same
parameters as an XML::XPath object constructor to initialize the primary
XML document object (the object which subsequent XML documents will be
merged into). These parameters can be any one of:
'filename' => 'SomeFile.xml'
'xml' => $variable_which_holds_a_bunch_of_XML_data
'ioref' => $file_InputOutput_reference
'context' => $existing_node_at_specified_context_to_become_new_obj
Merge's new() can also accept merge-option parameters to override the
default merge behavior. These include:
'conflict_resolution_method' => 'main', # main file wins
'conflict_resolution_method' => 'merg', # merge file wins
# 'last-in_wins' is the same as 'merg'
'conflict_resolution_method' => 'warn', # croak conflicts
'conflict_resolution_method' => 'test', # just test, 1 if conflict
merge()
The merge() member function can accept the same XML::XPath constructor
options as new() but this time they are for the temporary file which
will be merged into the main object. Merge-options from new() can also
be specified and they will only impact one particular invokation of
merge(). The specified document will be merged into the primary XML
document object according to the following default merge rules:
1. If both documents share the same root element name, they are merged
directly.
2. If they don't share root elements but the temporary merge file's root
element is found anywhere within the main file, the merge occurs at the
match.
3. If no root element match is found, the merge document becomes the new
last child of the main file's root element.
4. Whenever a deeper level is found with an element of the same name in
both documents and either it does not contain any distinguishing
attributes or it has attributes which are recognized as 'identifier'
(id) attributes (by default, for any element, these are attributes
named: 'id', 'idx', 'ndx', 'index', 'name', and 'handle'), a
corresponding element is searched for to match and merge with.
5. Any remaining (non-id) nodes are merged in document order.
6. When a conflict arises as non-id attributes or other nodes merge, the
specified conflict_resolution_method merge-option is applied (which by
default has the main file data persist at the expense of the merging
file data).
Some of the above rules can be overridden first by the object's
merge-options and second by the particular method call's merge-options.
Thus, if the default merge-option for conflict resolution is to have the
main object win and you use the following constructor:
my $merge_obj = XML::Merge->new(
'filename' => 'MainFile.xml',
'conflict_resolution_method' => 'last-in_wins');
... then any $merge_obj->merge() call would override the default merge
behavior by letting the document being merged have priority over the
main object's document. However, you could supply additional
merge-options in the parameter list of your specific merge() call like:
$merge_obj->merge(
'filename' => 'File2Add.xml',
'conflict_resolution_method' => 'warn');
... to have the latest option override further.
The 'test' conflict_resolution_method merge-option does not modify the
object at all. It solely returns zero (0) if no conflict was encountered
from a temporary attempted merge.
It should be used like:
for(@files) {
if($merge_obj->merge('cres' => 'test', $_)) {
croak("Yipes! Conflict with file:$_!\n");
} else {
$merge_obj->merge($_); # only do it if there are no conflicts
}
}
merge() can also accept another XML::Merge object as a parameter for
what to be merged with the main object instead of a filename. An example
of this is:
$merge_obj->merge($another_merge_obj);
Along with the merge options that can be specified in the object
constructor, merge() also accepts the following options to specify where
to perform the merge relative to:
'merge_destination_path' => $main_obj_xpath_location,
'merge_source_path' => $merging_obj_xpath_location,
unmerge()
The unmerge() member function is a shorthand for calling both write()
and prune() on a certain XPath location which should be written out to a
disk file before being removed from the Merge object. Please see
XML::Tidy for documentation of the inherited write() and prune() member
functions.
This unmerge() process could be the opposite of merge() if no original
elements or attributes overlapped and combined but if combining did
happen, this would remove original sections of your primary XML
document's data from your Merge object so please use this carefully. It
is meant to help separate a giant object (probably the result of myriad
merge() calls) back into separate useful well-formed XML documents on
disk.
unmerge() takes a filename and an xpath_location parameter.
Accessors
get_object_to_merge()
Returns the object which was last merged into the main object.
set_object_to_merge()
Assigns the object which was last merged into the main object.
get_conflict_resolution_method()
Returns the underlying merge-option conflict_resolution_method.
set_conflict_resolution_method()
A new value can be provided as a parameter to be assigned as the
XML::Merge object's merge-option.
get_id_xpath_list()
Returns the underlying id_xpath_list. This is normally just a list of
attributes (e.g., '@id', '@idx', '@ndx', '@index', '@name', '@handle')
which are unique identifiers for any XML element within merging instance
documents. When these attribute names are encountered during a merge(),
another element with the same name and attribute value are searched for
explicitly in order to align deeper merging and conflict resolution.
set_id_xpath_list()
A new list can assigned to the XML::Merge object's id_xpath_list.
Please note that this list normally contains XPath attributes so they
must be preceded by an at-symbol (@) like: '@example_new_id_attribute'.
CHANGES
Revision history for Perl extension XML::Merge:
- 1.4 G7NMEdxm Sat Jul 23 14:39:59:48 -0500 2016
* inverted conflict resolution 'test' value since true 1 for conflict
makes more sense
* renumbered t/*.t
* updated Makefile.PL and Build.PL to hopefully fix issue
<HTTPS://RT.CPAN.Org/Public/Bug/Display.html?id=29898> (Thanks Kevin.)
* removed DBUG printing
* removed PT from VERSION to fix issue
<HTTPS://RT.CPAN.Org/Public/Bug/Display.html?id=106873> (Thanks
ppisar.)
* updated license to GPLv3
- 1.2.75BAJNl Fri May 11 10:19:23:47 2007
* added default id @s: idx, ndx, and index
- 1.2.565EgGd Sun Jun 5 14:42:16:39 2005
* added use XML::Tidy to make sure exports are available
* removed 02prune.t and moved 03keep.t to 02keep.t ... passing tests
is good
- 1.2.4CCJWiB Sun Dec 12 19:32:44:11 2004
* guessing how to fix Darwin test failure @ t/02prune.t first prune()
call
- 1.0.4CAL5IS Fri Dec 10 21:05:18:28 2004
* fixed buggy _recmerge
- 1.0.4CAEU0I Fri Dec 10 14:30:00:18 2004
* made accessors for _id_xpath_list
* made _id_xpath_list take XPath locations instead of elem names (old
_idea)
* made test _cres (at Marc's request)
* made warn _cres croak
* made Merge inherit from Tidy (which inherits from XPath)
* separated reload(), strip(), tidy(), prune(), and write() into own
XML::Tidy module
- 1.0.4C2Nf0R Thu Dec 2 23:41:00:27 2004
* updated license and prep'd for release
- 1.0.4C2BcI2 Thu Dec 2 11:38:18:02 2004
* updated reload(), strip(), and tidy() to verify _xpob exists
- 1.0.4C1JHOl Wed Dec 1 19:17:24:47 2004
* commented out override stuff since it's probably bad form and dumps
crap warnings all over tests and causes them to fail... so I guess
just uncomment that stuff if you care to preserve PI's and escapes
- 1.0.4C1J7gt Wed Dec 1 19:07:42:55 2004
* made merge() accept merge_source_xpath and merge_destination_xpath
params
* made merge() accept other Merge objects
* made reload() not clobber basic escapes (by overriding Text
toString())
* made tidy() not kill processing-instructions (by overriding
node_test())
* made tidy() not kill comments
- 1.0.4BOHGjm Wed Nov 24 17:16:45:48 2004
* fixed merge() same elems with diff ids bug
- 1.0.4BNBCZL Tue Nov 23 11:12:35:21 2004
* rewrote both merge() and _recmerge() _cres stuff since it was buggy
before... so hopefully consistently good now
- 1.0.4BMJCPm Mon Nov 22 19:12:25:48 2004
* fixed merge() for empty elem matching and _cres on text kids
- 1.0.4BMGTLF Mon Nov 22 16:29:21:15 2004
* separated reload() from strip() so that prune() can call it too
- 1.0.4BM0B3x Mon Nov 22 00:11:03:59 2004
* fixed tidy() empty elem bug and implemented prune() and unmerge()
- 1.0.4BJAZpM Fri Nov 19 10:35:51:22 2004
* fixing e() ABSTRACT gen bug
- 1.0.4BJAMR6 Fri Nov 19 10:22:27:06 2004
* fleshed out POD and members
- 1.0.4AIDqmR Mon Oct 18 13:52:48:27 2004
* original version
TODO
- add Kevin's multiple _idea option where several element attributes are
an ID together, from:
<HTTPS://RT.CPAN.Org/Public/Bug/Display.html?id=29897>
- make namespaces and attributes stay in order after merge()
- make text append merge option
- handle comment joins and stamping options
- support modification-time conflict resolution method
- add _ignr ignore list of merge XPath locations to not merge
(pre-prune())
INSTALL
From the command shell, please run:
`perl -MCPAN -e "install XML::Merge"`
or uncompress the package and run the standard:
`perl Makefile.PL; make; make test; make install`
or if you don't have `make` but Module::Build is installed, try:
`perl Build.PL; perl Build; perl Build test; perl Build install`
FILES
XML::Merge requires:
Carp to allow errors to croak() from calling sub
XML::Tidy to use objects derived from XPath to update XML
LICENSE
Most source code should be Free! Code I have lawful authority over is
and shall be! Copyright: (c) 2004-2016, Pip Stuart. Copyleft : This
software is licensed under the GNU General Public License (version 3 or
later). Please consult <HTTP://GNU.Org/licenses/gpl-3.0.txt> for
important information about your freedom. This is Free Software: you are
free to change and redistribute it. There is NO WARRANTY, to the extent
permitted by law. See <HTTP://FSF.Org> for further information.
AUTHOR
Pip Stuart <Pip@CPAN.Org>