XML documents in a relational database.

depage-xmldb stores XML documents in MySQL/MariaDB, lets you query them with a small and fast XPath dialect and modify them node by node — without ever loading the whole document into memory.

It is the XML storage layer of depage-cms, but it works just as well as a standalone library.

Features

XML documents in a relational database. Every document is a tree of nodes in a _xmltree table, the document metadata lives in _xmldocs.
XPath queries. A small XPath dialect is translated into SQL and answered directly by the database. Queries that the SQL path cannot express fall back to a DOM XPath evaluation.
Incremental saves. savePatch() and saveNodePatch() only write the subtrees that really changed. Merkle hashes over the tree make it possible to find them without comparing every single node.
Stable node ids. Every node carries a db:id, so a node can be referenced, addressed and updated without knowing its position in the document.
Doctype handlers. Structure rules, permissions, whitespace handling and change hooks are implemented by a handler class of your choice.
Version history. Every document keeps its versions, can publish them and restore older ones.
Transactions. Nested transactions with savepoints, so a failed change can be rolled back without losing the surrounding work.
Batched writes. Rows are written with multi-row INSERT and UPDATE ... CASE statements, independent of the size of the document.
Caching. Subdocuments and query results are cached per document and the whole cache of a document is dropped in one go when it changes.

Requirements

PHP >= 8.3 with the extensions dom, pdo and pdo_mysql
MySQL >= 8.0 or MariaDB >= 10.2 (the library uses recursive common table expressions and window functions)
the libraries depage/db, depage/xml and depage/cache

Installation

composer require depage/xmldb

Getting started

use Depage\Cache\Cache;
use Depage\Db\Pdo;
use Depage\XmlDb\XmlDb;
 
$pdo = new Pdo('mysql:host=localhost;dbname=depage', 'user', 'password', ['prefix' => 'dp_']);
$cache = Cache::factory('xmldb', ['disposition' => 'file', 'cachepath' => '/var/cache/depage']);
 
$xmlDb = new XmlDb('dp_', $pdo, $cache);
$xmlDb->updateSchema();          // create or migrate the tables
 
// a document is addressed by its name or by its id
$doc = $xmlDb->createDoc(\Depage\XmlDb\XmlDoctypes\Base::class, 'pages');
 
$dom = new \Depage\Xml\Document();
$dom->loadXML(
    '<dpg:pages xmlns:dpg="http://www.depagecms.net/ns/depage">'
    . '<pg:page name="home" xmlns:pg="http://www.depagecms.net/ns/page"/>'
    . '</dpg:pages>'
);
$doc->save($dom);
 
// read the document back (without the db:id attributes)
$xml = $doc->getXml(false);
 
// query it
$nodeIds = $doc->getNodeIdsByXpath('/dpg:pages/pg:page[@name="home"]');
 
// a node is addressed by its db:id
$page = $doc->getSubdocByNodeId($nodeIds[0]);

Writing to a document is done either as a whole or node by node:

// full save: the document is rewritten
$doc->save($dom);
 
// incremental save: only the changed subtrees are written
$doc->savePatch($dom);
print_r($doc->getLastPatchStats());
 
// node by node
$doc->addNode($domNode, $targetNodeId);
$doc->moveNodeAfter($nodeId, $targetNodeId);
$doc->setAttribute($nodeId, 'name', 'home');

Documentation

Getting started — setup, documents, reading, writing
XPath reference — the supported XPath dialect
Doctype handlers — structure rules, permissions, hooks
Version history — versions, publishing, restoring
Storage internals — tables, positions, hashing, batching
Developer guide — tests, code style, documentation

The API reference is generated with Doxygen and published at https://docs.depage.net/depage-xmldb/.

Development

make test       # run the PHPUnit test suite (needs a MySQL/MariaDB server)
make cs         # check the code style
make compat     # check the PHP version compatibility
make doc        # generate the API documentation

License (dual)

GPL-2.0: https://www.gnu.org/licenses/gpl-2.0.html
MIT: https://opensource.org/licenses/MIT