<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en"> <head> <meta http-equiv="Content-Type" content="text/html; charset=utf-8" /> <meta name="generator" content="Docutils 0.10: http://docutils.sourceforge.net/" /> <title>Benchmarks and Speed</title> <meta name="author" content="Stefan Behnel" /> <meta content="Performance evaluation of lxml and ElementTree: fast operations, common pitfalls and optimisation hints." name="description" /> <meta content="Python XML parser performance, XML processing, performance comparison, lxml performance, lxml.etree, lxml.objectify, benchmarks, ElementTree" name="keywords" /> <link rel="stylesheet" href="style.css" type="text/css" /> </head> <body> <div class="document" id="benchmarks-and-speed"> <div class="sidemenu"><ul id="lxml-section"><li><span class="section title">lxml</span><ul class="menu foreign" id="index-menu"><li class="menu title"><a href="index.html">lxml</a><ul class="submenu"><li class="menu item"><a href="index.html#introduction">Introduction</a></li><li class="menu item"><a href="index.html#support-the-project">Support the project</a></li><li class="menu item"><a href="index.html#documentation">Documentation</a></li><li class="menu item"><a href="index.html#download">Download</a></li><li class="menu item"><a href="index.html#mailing-list">Mailing list</a></li><li class="menu item"><a href="index.html#bug-tracker">Bug tracker</a></li><li class="menu item"><a href="index.html#license">License</a></li><li class="menu item"><a href="index.html#old-versions">Old Versions</a></li><li class="menu item"><a href="index.html#legal-notice-for-donations">Legal Notice for Donations</a></li></ul></li></ul><ul class="menu foreign" id="intro-menu"><li class="menu title"><a href="intro.html">Why lxml?</a><ul class="submenu"><li class="menu item"><a href="intro.html#motto">Motto</a></li><li class="menu item"><a href="intro.html#aims">Aims</a></li></ul></li></ul><ul class="menu foreign" id="installation-menu"><li class="menu title"><a href="installation.html">Installing lxml</a><ul class="submenu"><li class="menu item"><a href="installation.html#requirements">Requirements</a></li><li class="menu item"><a href="installation.html#installation">Installation</a></li><li class="menu item"><a href="installation.html#building-lxml-from-sources">Building lxml from sources</a></li><li class="menu item"><a href="installation.html#using-lxml-with-python-libxml2">Using lxml with python-libxml2</a></li><li class="menu item"><a href="installation.html#ms-windows">MS Windows</a></li><li class="menu item"><a href="installation.html#macos-x">MacOS-X</a></li></ul></li></ul><ul class="menu current" id="performance-menu"><li class="menu title"><a href="performance.html">Benchmarks and Speed</a><ul class="submenu"><li class="menu item"><a href="performance.html#general-notes">General notes</a></li><li class="menu item"><a href="performance.html#how-to-read-the-timings">How to read the timings</a></li><li class="menu item"><a href="performance.html#parsing-and-serialising">Parsing and Serialising</a></li><li class="menu item"><a href="performance.html#the-elementtree-api">The ElementTree API</a></li><li class="menu item"><a href="performance.html#xpath">XPath</a></li><li class="menu item"><a href="performance.html#a-longer-example">A longer example</a></li><li class="menu item"><a href="performance.html#lxml-objectify">lxml.objectify</a></li></ul></li></ul><ul class="menu foreign" id="compatibility-menu"><li class="menu title"><a href="compatibility.html">ElementTree compatibility of lxml.etree</a></li></ul><ul class="menu foreign" id="FAQ-menu"><li class="menu title"><a href="FAQ.html">lxml FAQ - Frequently Asked Questions</a><ul class="submenu"><li class="menu item"><a href="FAQ.html#general-questions">General Questions</a></li><li class="menu item"><a href="FAQ.html#installation">Installation</a></li><li class="menu item"><a href="FAQ.html#contributing">Contributing</a></li><li class="menu item"><a href="FAQ.html#bugs">Bugs</a></li><li class="menu item"><a href="FAQ.html#id1">Threading</a></li><li class="menu item"><a href="FAQ.html#parsing-and-serialisation">Parsing and Serialisation</a></li><li class="menu item"><a href="FAQ.html#xpath-and-document-traversal">XPath and Document Traversal</a></li></ul></li></ul></li></ul><ul id="Developing with lxml-section"><li><span class="section title">Developing with lxml</span><ul class="menu foreign" id="tutorial-menu"><li class="menu title"><a href="tutorial.html">The lxml.etree Tutorial</a><ul class="submenu"><li class="menu item"><a href="tutorial.html#the-element-class">The Element class</a></li><li class="menu item"><a href="tutorial.html#the-elementtree-class">The ElementTree class</a></li><li class="menu item"><a href="tutorial.html#parsing-from-strings-and-files">Parsing from strings and files</a></li><li class="menu item"><a href="tutorial.html#namespaces">Namespaces</a></li><li class="menu item"><a href="tutorial.html#the-e-factory">The E-factory</a></li><li class="menu item"><a href="tutorial.html#elementpath">ElementPath</a></li></ul></li></ul><ul class="menu foreign" id="api index-menu"><li class="menu title"><a href="api/index.html">API reference</a></li></ul><ul class="menu foreign" id="api-menu"><li class="menu title"><a href="api.html">APIs specific to lxml.etree</a><ul class="submenu"><li class="menu item"><a href="api.html#lxml-etree">lxml.etree</a></li><li class="menu item"><a href="api.html#other-element-apis">Other Element APIs</a></li><li class="menu item"><a href="api.html#trees-and-documents">Trees and Documents</a></li><li class="menu item"><a href="api.html#iteration">Iteration</a></li><li class="menu item"><a href="api.html#error-handling-on-exceptions">Error handling on exceptions</a></li><li class="menu item"><a href="api.html#error-logging">Error logging</a></li><li class="menu item"><a href="api.html#serialisation">Serialisation</a></li><li class="menu item"><a href="api.html#incremental-xml-generation">Incremental XML generation</a></li><li class="menu item"><a href="api.html#cdata">CDATA</a></li><li class="menu item"><a href="api.html#xinclude-and-elementinclude">XInclude and ElementInclude</a></li><li class="menu item"><a href="api.html#write-c14n-on-elementtree">write_c14n on ElementTree</a></li></ul></li></ul><ul class="menu foreign" id="parsing-menu"><li class="menu title"><a href="parsing.html">Parsing XML and HTML with lxml</a><ul class="submenu"><li class="menu item"><a href="parsing.html#parsers">Parsers</a></li><li class="menu item"><a href="parsing.html#the-target-parser-interface">The target parser interface</a></li><li class="menu item"><a href="parsing.html#the-feed-parser-interface">The feed parser interface</a></li><li class="menu item"><a href="parsing.html#iterparse-and-iterwalk">iterparse and iterwalk</a></li><li class="menu item"><a href="parsing.html#python-unicode-strings">Python unicode strings</a></li></ul></li></ul><ul class="menu foreign" id="validation-menu"><li class="menu title"><a href="validation.html">Validation with lxml</a><ul class="submenu"><li class="menu item"><a href="validation.html#validation-at-parse-time">Validation at parse time</a></li><li class="menu item"><a href="validation.html#id1">DTD</a></li><li class="menu item"><a href="validation.html#relaxng">RelaxNG</a></li><li class="menu item"><a href="validation.html#xmlschema">XMLSchema</a></li><li class="menu item"><a href="validation.html#id2">Schematron</a></li><li class="menu item"><a href="validation.html#id3">(Pre-ISO-Schematron)</a></li></ul></li></ul><ul class="menu foreign" id="xpathxslt-menu"><li class="menu title"><a href="xpathxslt.html">XPath and XSLT with lxml</a><ul class="submenu"><li class="menu item"><a href="xpathxslt.html#xpath">XPath</a></li><li class="menu item"><a href="xpathxslt.html#xslt">XSLT</a></li></ul></li></ul><ul class="menu foreign" id="objectify-menu"><li class="menu title"><a href="objectify.html">lxml.objectify</a><ul class="submenu"><li class="menu item"><a href="objectify.html#the-lxml-objectify-api">The lxml.objectify API</a></li><li class="menu item"><a href="objectify.html#asserting-a-schema">Asserting a Schema</a></li><li class="menu item"><a href="objectify.html#objectpath">ObjectPath</a></li><li class="menu item"><a href="objectify.html#python-data-types">Python data types</a></li><li class="menu item"><a href="objectify.html#how-data-types-are-matched">How data types are matched</a></li><li class="menu item"><a href="objectify.html#what-is-different-from-lxml-etree">What is different from lxml.etree?</a></li></ul></li></ul><ul class="menu foreign" id="lxmlhtml-menu"><li class="menu title"><a href="lxmlhtml.html">lxml.html</a><ul class="submenu"><li class="menu item"><a href="lxmlhtml.html#parsing-html">Parsing HTML</a></li><li class="menu item"><a href="lxmlhtml.html#html-element-methods">HTML Element Methods</a></li><li class="menu item"><a href="lxmlhtml.html#running-html-doctests">Running HTML doctests</a></li><li class="menu item"><a href="lxmlhtml.html#creating-html-with-the-e-factory">Creating HTML with the E-factory</a></li><li class="menu item"><a href="lxmlhtml.html#working-with-links">Working with links</a></li><li class="menu item"><a href="lxmlhtml.html#forms">Forms</a></li><li class="menu item"><a href="lxmlhtml.html#cleaning-up-html">Cleaning up HTML</a></li><li class="menu item"><a href="lxmlhtml.html#html-diff">HTML Diff</a></li><li class="menu item"><a href="lxmlhtml.html#examples">Examples</a></li></ul></li></ul><ul class="menu foreign" id="cssselect-menu"><li class="menu title"><a href="cssselect.html">lxml.cssselect</a><ul class="submenu"><li class="menu item"><a href="cssselect.html#the-cssselector-class">The CSSSelector class</a></li><li class="menu item"><a href="cssselect.html#the-cssselect-method">The cssselect method</a></li><li class="menu item"><a href="cssselect.html#supported-selectors">Supported Selectors</a></li><li class="menu item"><a href="cssselect.html#namespaces">Namespaces</a></li></ul></li></ul><ul class="menu foreign" id="elementsoup-menu"><li class="menu title"><a href="elementsoup.html">BeautifulSoup Parser</a><ul class="submenu"><li class="menu item"><a href="elementsoup.html#parsing-with-the-soupparser">Parsing with the soupparser</a></li><li class="menu item"><a href="elementsoup.html#entity-handling">Entity handling</a></li><li class="menu item"><a href="elementsoup.html#using-soupparser-as-a-fallback">Using soupparser as a fallback</a></li><li class="menu item"><a href="elementsoup.html#using-only-the-encoding-detection">Using only the encoding detection</a></li></ul></li></ul><ul class="menu foreign" id="html5parser-menu"><li class="menu title"><a href="html5parser.html">html5lib Parser</a><ul class="submenu"><li class="menu item"><a href="html5parser.html#differences-to-regular-html-parsing">Differences to regular HTML parsing</a></li><li class="menu item"><a href="html5parser.html#function-reference">Function Reference</a></li></ul></li></ul></li></ul><ul id="Extending lxml-section"><li><span class="section title">Extending lxml</span><ul class="menu foreign" id="resolvers-menu"><li class="menu title"><a href="resolvers.html">Document loading and URL resolving</a><ul class="submenu"><li class="menu item"><a href="resolvers.html#xml-catalogs">XML Catalogs</a></li><li class="menu item"><a href="resolvers.html#uri-resolvers">URI Resolvers</a></li><li class="menu item"><a href="resolvers.html#document-loading-in-context">Document loading in context</a></li><li class="menu item"><a href="resolvers.html#i-o-access-control-in-xslt">I/O access control in XSLT</a></li></ul></li></ul><ul class="menu foreign" id="extensions-menu"><li class="menu title"><a href="extensions.html">Python extensions for XPath and XSLT</a><ul class="submenu"><li class="menu item"><a href="extensions.html#xpath-extension-functions">XPath Extension functions</a></li><li class="menu item"><a href="extensions.html#xslt-extension-elements">XSLT extension elements</a></li></ul></li></ul><ul class="menu foreign" id="element classes-menu"><li class="menu title"><a href="element_classes.html">Using custom Element classes in lxml</a><ul class="submenu"><li class="menu item"><a href="element_classes.html#background-on-element-proxies">Background on Element proxies</a></li><li class="menu item"><a href="element_classes.html#element-initialization">Element initialization</a></li><li class="menu item"><a href="element_classes.html#setting-up-a-class-lookup-scheme">Setting up a class lookup scheme</a></li><li class="menu item"><a href="element_classes.html#generating-xml-with-custom-classes">Generating XML with custom classes</a></li><li class="menu item"><a href="element_classes.html#id1">Implementing namespaces</a></li></ul></li></ul><ul class="menu foreign" id="sax-menu"><li class="menu title"><a href="sax.html">Sax support</a><ul class="submenu"><li class="menu item"><a href="sax.html#building-a-tree-from-sax-events">Building a tree from SAX events</a></li><li class="menu item"><a href="sax.html#producing-sax-events-from-an-elementtree-or-element">Producing SAX events from an ElementTree or Element</a></li><li class="menu item"><a href="sax.html#interfacing-with-pulldom-minidom">Interfacing with pulldom/minidom</a></li></ul></li></ul><ul class="menu foreign" id="capi-menu"><li class="menu title"><a href="capi.html">The public C-API of lxml.etree</a><ul class="submenu"><li class="menu item"><a href="capi.html#writing-external-modules-in-cython">Writing external modules in Cython</a></li><li class="menu item"><a href="capi.html#writing-external-modules-in-c">Writing external modules in C</a></li></ul></li></ul></li></ul><ul id="Developing lxml-section"><li><span class="section title">Developing lxml</span><ul class="menu foreign" id="build-menu"><li class="menu title"><a href="build.html">How to build lxml from source</a><ul class="submenu"><li class="menu item"><a href="build.html#cython">Cython</a></li><li class="menu item"><a href="build.html#github-git-and-hg">Github, git and hg</a></li><li class="menu item"><a href="build.html#building-the-sources">Building the sources</a></li><li class="menu item"><a href="build.html#running-the-tests-and-reporting-errors">Running the tests and reporting errors</a></li><li class="menu item"><a href="build.html#building-an-egg">Building an egg</a></li><li class="menu item"><a href="build.html#building-lxml-on-macos-x">Building lxml on MacOS-X</a></li><li class="menu item"><a href="build.html#static-linking-on-windows">Static linking on Windows</a></li><li class="menu item"><a href="build.html#building-debian-packages-from-svn-sources">Building Debian packages from SVN sources</a></li></ul></li></ul><ul class="menu foreign" id="lxml source howto-menu"><li class="menu title"><a href="lxml-source-howto.html">How to read the source of lxml</a><ul class="submenu"><li class="menu item"><a href="lxml-source-howto.html#what-is-cython">What is Cython?</a></li><li class="menu item"><a href="lxml-source-howto.html#where-to-start">Where to start?</a></li><li class="menu item"><a href="lxml-source-howto.html#lxml-etree">lxml.etree</a></li><li class="menu item"><a href="lxml-source-howto.html#python-modules">Python modules</a></li><li class="menu item"><a href="lxml-source-howto.html#lxml-objectify">lxml.objectify</a></li><li class="menu item"><a href="lxml-source-howto.html#lxml-html">lxml.html</a></li></ul></li></ul><ul class="menu foreign" id="changes 3 2 4-menu"><li class="menu title"><a href="changes-3.2.4.html">Release Changelog</a></li></ul><ul class="menu foreign" id="credits-menu"><li class="menu title"><a href="credits.html">Credits</a><ul class="submenu"><li class="menu item"><a href="credits.html#main-contributors">Main contributors</a></li><li class="menu item"><a href="credits.html#special-thanks-goes-to">Special thanks goes to:</a></li></ul></li></ul></li><li><a href="http://lxml.de/sitemap.html">Sitemap</a></li></ul></div><h1 class="title">Benchmarks and Speed</h1> <table class="docinfo" frame="void" rules="none"> <col class="docinfo-name" /> <col class="docinfo-content" /> <tbody valign="top"> <tr><th class="docinfo-name">Author:</th> <td>Stefan Behnel</td></tr> </tbody> </table> <p>lxml.etree is a very fast XML library. Most of this is due to the speed of libxml2, e.g. the parser and serialiser, or the XPath engine. Other areas of lxml were specifically written for high performance in high-level operations, such as the tree iterators.</p> <p>On the other hand, the simplicity of lxml sometimes hides internal operations that are more costly than the API suggests. If you are not aware of these cases, lxml may not always perform as you expect. A common example in the Python world is the Python list type. New users often expect it to be a linked list, while it actually is implemented as an array, which results in a completely different complexity for common operations.</p> <p>Similarly, the tree model of libxml2 is more complex than what lxml's ElementTree API projects into Python space, so some operations may show unexpected performance. Rest assured that most lxml users will not notice this in real life, as lxml is very fast in absolute numbers. It is definitely fast enough for most applications, so lxml is probably somewhere between 'fast enough' and 'the best choice' for yours. Read some <a class="reference external" href="http://permalink.gmane.org/gmane.comp.python.lxml.devel/3250">messages</a> from <a class="reference external" href="http://article.gmane.org/gmane.comp.python.lxml.devel/3246">happy</a> <a class="reference external" href="http://thread.gmane.org/gmane.comp.python.lxml.devel/3244/focus=3244">users</a> to see what we mean.</p> <p>This text describes where lxml.etree (abbreviated to 'lxe') excels, gives hints on some performance traps and compares the overall performance to the original <a class="reference external" href="http://effbot.org/zone/element-index.htm">ElementTree</a> (ET) and <a class="reference external" href="http://effbot.org/zone/celementtree.htm">cElementTree</a> (cET) libraries by Fredrik Lundh. The cElementTree library is a fast C-implementation of the original ElementTree.</p> <div class="contents topic" id="contents"> <p class="topic-title first">Contents</p> <ul class="simple"> <li><a class="reference internal" href="#general-notes" id="id1">General notes</a></li> <li><a class="reference internal" href="#how-to-read-the-timings" id="id2">How to read the timings</a></li> <li><a class="reference internal" href="#parsing-and-serialising" id="id3">Parsing and Serialising</a></li> <li><a class="reference internal" href="#the-elementtree-api" id="id4">The ElementTree API</a><ul> <li><a class="reference internal" href="#child-access" id="id5">Child access</a></li> <li><a class="reference internal" href="#element-creation" id="id6">Element creation</a></li> <li><a class="reference internal" href="#merging-different-sources" id="id7">Merging different sources</a></li> <li><a class="reference internal" href="#deepcopy" id="id8">deepcopy</a></li> <li><a class="reference internal" href="#tree-traversal" id="id9">Tree traversal</a></li> </ul> </li> <li><a class="reference internal" href="#xpath" id="id10">XPath</a></li> <li><a class="reference internal" href="#a-longer-example" id="id11">A longer example</a></li> <li><a class="reference internal" href="#lxml-objectify" id="id12">lxml.objectify</a><ul> <li><a class="reference internal" href="#objectpath" id="id13">ObjectPath</a></li> <li><a class="reference internal" href="#caching-elements" id="id14">Caching Elements</a></li> <li><a class="reference internal" href="#further-optimisations" id="id15">Further optimisations</a></li> </ul> </li> </ul> </div> <div class="section" id="general-notes"> <h1>General notes</h1> <p>First thing to say: there <em>is</em> an overhead involved in having a DOM-like C library mimic the ElementTree API. As opposed to ElementTree, lxml has to generate Python representations of tree nodes on the fly when asked for them, and the internal tree structure of libxml2 results in a higher maintenance overhead than the simpler top-down structure of ElementTree. What this means is: the more of your code runs in Python, the less you can benefit from the speed of lxml and libxml2. Note, however, that this is true for most performance critical Python applications. No one would implement fourier transformations in pure Python when you can use NumPy.</p> <p>The up side then is that lxml provides powerful tools like tree iterators, XPath and XSLT, that can handle complex operations at the speed of C. Their pythonic API in lxml makes them so flexible that most applications can easily benefit from them.</p> </div> <div class="section" id="how-to-read-the-timings"> <h1>How to read the timings</h1> <p>The statements made here are backed by the (micro-)benchmark scripts <a class="reference external" href="https://github.com/lxml/lxml/blob/master/benchmark/bench_etree.py">bench_etree.py</a>, <a class="reference external" href="https://github.com/lxml/lxml/blob/master/benchmark/bench_xpath.py">bench_xpath.py</a> and <a class="reference external" href="https://github.com/lxml/lxml/blob/master/benchmark/bench_objectify.py">bench_objectify.py</a> that come with the lxml source distribution. They are distributed under the same BSD license as lxml itself, and the lxml project would like to promote them as a general benchmarking suite for all ElementTree implementations. New benchmarks are very easy to add as tiny test methods, so if you write a performance test for a specific part of the API yourself, please consider sending it to the lxml mailing list.</p> <p>The timings presented below compare lxml 3.1.1 (with libxml2 2.9.0) to the latest released versions of ElementTree (with cElementTree as accelerator module) in the standard library of CPython 3.3.0. They were run single-threaded on a 2.9GHz 64bit double core Intel i7 machine under Ubuntu Linux 12.10 (Quantal). The C libraries were compiled with the same platform specific optimisation flags. The Python interpreter was also manually compiled for the platform. Note that many of the following ElementTree timings are therefore better than what a normal Python installation with the standard library (c)ElementTree modules would yield. Note also that CPython 2.7 and 3.2+ come with a newer ElementTree version, so older Python installations will not perform as good for (c)ElementTree, and sometimes substantially worse.</p> <p>The scripts run a number of simple tests on the different libraries, using different XML tree configurations: different tree sizes (T1-4), with or without attributes (-/A), with or without ASCII string or unicode text (-/S/U), and either against a tree or its serialised XML form (T/X). In the result extracts cited below, T1 refers to a 3-level tree with many children at the third level, T2 is swapped around to have many children below the root element, T3 is a deep tree with few children at each level and T4 is a small tree, slightly broader than deep. If repetition is involved, this usually means running the benchmark in a loop over all children of the tree root, otherwise, the operation is run on the root node (C/R).</p> <p>As an example, the character code <tt class="docutils literal">(SATR T1)</tt> states that the benchmark was running for tree T1, with plain string text (S) and attributes (A). It was run against the root element (R) in the tree structure of the data (T).</p> <p>Note that very small operations are repeated in integer loops to make them measurable. It is therefore not always possible to compare the absolute timings of, say, a single access benchmark (which usually loops) and a 'get all in one step' benchmark, which already takes enough time to be measurable and is therefore measured as is. An example is the index access to a single child, which cannot be compared to the timings for <tt class="docutils literal">getchildren()</tt>. Take a look at the concrete benchmarks in the scripts to understand how the numbers compare.</p> </div> <div class="section" id="parsing-and-serialising"> <h1>Parsing and Serialising</h1> <p>Serialisation is an area where lxml excels. The reason is that it executes entirely at the C level, without any interaction with Python code. The results are rather impressive, especially for UTF-8, which is native to libxml2. While 20 to 40 times faster than (c)ElementTree 1.2 (which was part of the standard library before Python 2.7/3.2), lxml is still more than 10 times as fast as the much improved ElementTree 1.3 in recent Python versions:</p> <pre class="literal-block"> lxe: tostring_utf16 (S-TR T1) 7.9958 msec/pass cET: tostring_utf16 (S-TR T1) 83.1358 msec/pass lxe: tostring_utf16 (UATR T1) 8.3222 msec/pass cET: tostring_utf16 (UATR T1) 84.4688 msec/pass lxe: tostring_utf16 (S-TR T2) 8.2297 msec/pass cET: tostring_utf16 (S-TR T2) 87.3415 msec/pass lxe: tostring_utf8 (S-TR T2) 6.5677 msec/pass cET: tostring_utf8 (S-TR T2) 76.2064 msec/pass lxe: tostring_utf8 (U-TR T3) 1.1952 msec/pass cET: tostring_utf8 (U-TR T3) 22.0058 msec/pass </pre> <p>The difference is somewhat smaller for plain text serialisation:</p> <pre class="literal-block"> lxe: tostring_text_ascii (S-TR T1) 2.7738 msec/pass cET: tostring_text_ascii (S-TR T1) 4.7629 msec/pass lxe: tostring_text_ascii (S-TR T3) 0.8273 msec/pass cET: tostring_text_ascii (S-TR T3) 1.5273 msec/pass lxe: tostring_text_utf16 (S-TR T1) 2.7659 msec/pass cET: tostring_text_utf16 (S-TR T1) 10.5038 msec/pass lxe: tostring_text_utf16 (U-TR T1) 2.8017 msec/pass cET: tostring_text_utf16 (U-TR T1) 10.5207 msec/pass </pre> <p>The <tt class="docutils literal">tostring()</tt> function also supports serialisation to a Python unicode string object, which is currently faster in ElementTree under CPython 3.3:</p> <pre class="literal-block"> lxe: tostring_text_unicode (S-TR T1) 2.6896 msec/pass cET: tostring_text_unicode (S-TR T1) 1.0056 msec/pass lxe: tostring_text_unicode (U-TR T1) 2.7366 msec/pass cET: tostring_text_unicode (U-TR T1) 1.0154 msec/pass lxe: tostring_text_unicode (S-TR T3) 0.7997 msec/pass cET: tostring_text_unicode (S-TR T3) 0.3154 msec/pass lxe: tostring_text_unicode (U-TR T4) 0.0048 msec/pass cET: tostring_text_unicode (U-TR T4) 0.0160 msec/pass </pre> <p>For parsing, lxml.etree and cElementTree compete for the medal. Depending on the input, either of the two can be faster. The (c)ET libraries use a very thin layer on top of the expat parser, which is known to be very fast. Here are some timings from the benchmarking suite:</p> <pre class="literal-block"> lxe: parse_bytesIO (SAXR T1) 13.0246 msec/pass cET: parse_bytesIO (SAXR T1) 8.2929 msec/pass lxe: parse_bytesIO (S-XR T3) 1.3542 msec/pass cET: parse_bytesIO (S-XR T3) 2.4023 msec/pass lxe: parse_bytesIO (UAXR T3) 7.5610 msec/pass cET: parse_bytesIO (UAXR T3) 11.2455 msec/pass </pre> <p>And another couple of timings <a class="reference external" href="http://svn.effbot.org/public/elementtree-1.3/benchmark.py">from a benchmark</a> that Fredrik Lundh <a class="reference external" href="http://effbot.org/zone/celementtree.htm#benchmarks">used to promote cElementTree</a>, comparing a number of different parsers. First, parsing a 274KB XML file containing Shakespeare's Hamlet:</p> <pre class="literal-block"> xml.etree.ElementTree.parse done in 0.017 seconds xml.etree.cElementTree.parse done in 0.007 seconds xml.etree.cElementTree.XMLParser.feed(): 6636 nodes read in 0.007 seconds lxml.etree.parse done in 0.003 seconds drop_whitespace.parse done in 0.003 seconds lxml.etree.XMLParser.feed(): 6636 nodes read in 0.004 seconds minidom tree read in 0.080 seconds </pre> <p>And a 3.4MB XML file containing the Old Testament:</p> <pre class="literal-block"> xml.etree.ElementTree.parse done in 0.038 seconds xml.etree.cElementTree.parse done in 0.030 seconds xml.etree.cElementTree.XMLParser.feed(): 25317 nodes read in 0.030 seconds lxml.etree.parse done in 0.016 seconds drop_whitespace.parse done in 0.015 seconds lxml.etree.XMLParser.feed(): 25317 nodes read in 0.022 seconds minidom tree read in 0.288 seconds </pre> <p>Here are the same benchmarks again, but including the memory usage of the process in KB before and after parsing (using os.fork() to make sure we start from a clean state each time). For the 274KB hamlet.xml file:</p> <pre class="literal-block"> Memory usage: 7284 xml.etree.ElementTree.parse done in 0.017 seconds Memory usage: 9432 (+2148) xml.etree.cElementTree.parse done in 0.007 seconds Memory usage: 9432 (+2152) xml.etree.cElementTree.XMLParser.feed(): 6636 nodes read in 0.007 seconds Memory usage: 9448 (+2164) lxml.etree.parse done in 0.003 seconds Memory usage: 11032 (+3748) drop_whitespace.parse done in 0.003 seconds Memory usage: 10224 (+2940) lxml.etree.XMLParser.feed(): 6636 nodes read in 0.004 seconds Memory usage: 11804 (+4520) minidom tree read in 0.080 seconds Memory usage: 12324 (+5040) </pre> <p>And for the 3.4MB Old Testament XML file:</p> <pre class="literal-block"> Memory usage: 10420 xml.etree.ElementTree.parse done in 0.038 seconds Memory usage: 20660 (+10240) xml.etree.cElementTree.parse done in 0.030 seconds Memory usage: 20660 (+10240) xml.etree.cElementTree.XMLParser.feed(): 25317 nodes read in 0.030 seconds Memory usage: 20844 (+10424) lxml.etree.parse done in 0.016 seconds Memory usage: 27624 (+17204) drop_whitespace.parse done in 0.015 seconds Memory usage: 24468 (+14052) lxml.etree.XMLParser.feed(): 25317 nodes read in 0.022 seconds Memory usage: 29844 (+19424) minidom tree read in 0.288 seconds Memory usage: 28788 (+18368) </pre> <p>As can be seen from the sizes, both lxml.etree and cElementTree are rather memory friendly compared to the pure Python libraries ElementTree and (especially) minidom. Comparing to older CPython versions, the memory footprint of the minidom library was considerably reduced in CPython 3.3, by about a factor of 4 in this case.</p> <p>For plain parser performance, lxml.etree and cElementTree tend to stay rather close to each other, usually within a factor of two, with winners well distributed over both sides. Similar timings can be observed for the <tt class="docutils literal">iterparse()</tt> function:</p> <pre class="literal-block"> lxe: iterparse_bytesIO (SAXR T1) 17.9198 msec/pass cET: iterparse_bytesIO (SAXR T1) 14.4982 msec/pass lxe: iterparse_bytesIO (UAXR T3) 8.8522 msec/pass cET: iterparse_bytesIO (UAXR T3) 12.9857 msec/pass </pre> <p>However, if you benchmark the complete round-trip of a serialise-parse cycle, the numbers will look similar to these:</p> <pre class="literal-block"> lxe: write_utf8_parse_bytesIO (S-TR T1) 19.8867 msec/pass cET: write_utf8_parse_bytesIO (S-TR T1) 80.7259 msec/pass lxe: write_utf8_parse_bytesIO (UATR T2) 23.7896 msec/pass cET: write_utf8_parse_bytesIO (UATR T2) 98.0766 msec/pass lxe: write_utf8_parse_bytesIO (S-TR T3) 3.0684 msec/pass cET: write_utf8_parse_bytesIO (S-TR T3) 24.6122 msec/pass lxe: write_utf8_parse_bytesIO (SATR T4) 0.3495 msec/pass cET: write_utf8_parse_bytesIO (SATR T4) 1.9610 msec/pass </pre> <p>For applications that require a high parser throughput of large files, and that do little to no serialization, both cET and lxml.etree are a good choice. The cET library is particularly fast for iterparse applications that extract small amounts of data or aggregate information from large XML data sets that do not fit into memory. If it comes to round-trip performance, however, lxml is multiple times faster in total. So, whenever the input documents are not considerably larger than the output, lxml is the clear winner.</p> <p>Regarding HTML parsing, Ian Bicking has done some <a class="reference external" href="http://blog.ianbicking.org/2008/03/30/python-html-parser-performance/">benchmarking on lxml's HTML parser</a>, comparing it to a number of other famous HTML parser tools for Python. lxml wins this contest by quite a length. To give an idea, the numbers suggest that lxml.html can run a couple of parse-serialise cycles in the time that other tools need for parsing alone. The comparison even shows some very favourable results regarding memory consumption.</p> <p>Liza Daly has written an article that presents a couple of tweaks to get the most out of lxml's parser for very large XML documents. She quite favourably positions <tt class="docutils literal">lxml.etree</tt> as a tool for <a class="reference external" href="http://www.ibm.com/developerworks/xml/library/x-hiperfparse/">high-performance XML parsing</a>.</p> <p>Finally, <a class="reference external" href="http://www.xml.com/">xml.com</a> has a couple of publications about XML parser performance. Farwick and Hafner have written two interesting articles that compare the parser of libxml2 to some major Java based XML parsers. One deals with <a class="reference external" href="http://www.xml.com/lpt/a/1702">event-driven parser performance</a>, the other one presents <a class="reference external" href="http://www.xml.com/lpt/a/1703">benchmark results comparing DOM parsers</a>. Both comparisons suggest that libxml2's parser performance is largely superiour to all commonly used Java parsers in almost all cases. Note that the C parser benchmark results are based on <a class="reference external" href="http://xmlbench.sourceforge.net/">xmlbench</a>, which uses a simpler setup for libxml2 than lxml does.</p> </div> <div class="section" id="the-elementtree-api"> <h1>The ElementTree API</h1> <p>Since all three libraries implement the same API, their performance is easy to compare in this area. A major disadvantage for lxml's performance is the different tree model that underlies libxml2. It allows lxml to provide parent pointers for elements and full XPath support, but also increases the overhead of tree building and restructuring. This can be seen from the tree setup times of the benchmark (given in seconds):</p> <pre class="literal-block"> lxe: -- S- U- -A SA UA T1: 0.0299 0.0343 0.0344 0.0293 0.0345 0.0342 T2: 0.0368 0.0423 0.0418 0.0427 0.0474 0.0459 T3: 0.0088 0.0084 0.0086 0.0251 0.0258 0.0261 T4: 0.0002 0.0002 0.0002 0.0005 0.0006 0.0006 cET: -- S- U- -A SA UA T1: 0.0050 0.0045 0.0093 0.0044 0.0043 0.0043 T2: 0.0073 0.0075 0.0074 0.0201 0.0075 0.0074 T3: 0.0033 0.0213 0.0032 0.0034 0.0033 0.0035 T4: 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 </pre> <p>The timings are somewhat close to each other, although cET can be several times faster than lxml for larger trees. One of the reasons is that lxml must encode incoming string data and tag names into UTF-8, and additionally discard the created Python elements after their use, when they are no longer referenced. ElementTree represents the tree itself through these objects, which reduces the overhead in creating them.</p> <div class="section" id="child-access"> <h2>Child access</h2> <p>The same tree overhead makes operations like collecting children as in <tt class="docutils literal">list(element)</tt> more costly in lxml. Where cET can quickly create a shallow copy of their list of children, lxml has to create a Python object for each child and collect them in a list:</p> <pre class="literal-block"> lxe: root_list_children (--TR T1) 0.0038 msec/pass cET: root_list_children (--TR T1) 0.0010 msec/pass lxe: root_list_children (--TR T2) 0.0455 msec/pass cET: root_list_children (--TR T2) 0.0050 msec/pass </pre> <p>This handicap is also visible when accessing single children:</p> <pre class="literal-block"> lxe: first_child (--TR T2) 0.0424 msec/pass cET: first_child (--TR T2) 0.0384 msec/pass lxe: last_child (--TR T1) 0.0477 msec/pass cET: last_child (--TR T1) 0.0467 msec/pass </pre> <p>... unless you also add the time to find a child index in a bigger list. ET and cET use Python lists here, which are based on arrays. The data structure used by libxml2 is a linked tree, and thus, a linked list of children:</p> <pre class="literal-block"> lxe: middle_child (--TR T1) 0.0710 msec/pass cET: middle_child (--TR T1) 0.0420 msec/pass lxe: middle_child (--TR T2) 1.7393 msec/pass cET: middle_child (--TR T2) 0.0396 msec/pass </pre> </div> <div class="section" id="element-creation"> <h2>Element creation</h2> <p>As opposed to ET, libxml2 has a notion of documents that each element must be in. This results in a major performance difference for creating independent Elements that end up in independently created documents:</p> <pre class="literal-block"> lxe: create_elements (--TC T2) 1.0045 msec/pass cET: create_elements (--TC T2) 0.0753 msec/pass </pre> <p>Therefore, it is always preferable to create Elements for the document they are supposed to end up in, either as SubElements of an Element or using the explicit <tt class="docutils literal">Element.makeelement()</tt> call:</p> <pre class="literal-block"> lxe: makeelement (--TC T2) 1.0586 msec/pass cET: makeelement (--TC T2) 0.1483 msec/pass lxe: create_subelements (--TC T2) 0.8826 msec/pass cET: create_subelements (--TC T2) 0.0827 msec/pass </pre> <p>So, if the main performance bottleneck of an application is creating large XML trees in memory through calls to Element and SubElement, cET is the best choice. Note, however, that the serialisation performance may even out this advantage, especially for smaller trees and trees with many attributes.</p> </div> <div class="section" id="merging-different-sources"> <h2>Merging different sources</h2> <p>A critical action for lxml is moving elements between document contexts. It requires lxml to do recursive adaptations throughout the moved tree structure.</p> <p>The following benchmark appends all root children of the second tree to the root of the first tree:</p> <pre class="literal-block"> lxe: append_from_document (--TR T1,T2) 1.0812 msec/pass cET: append_from_document (--TR T1,T2) 0.1104 msec/pass lxe: append_from_document (--TR T3,T4) 0.0155 msec/pass cET: append_from_document (--TR T3,T4) 0.0060 msec/pass </pre> <p>Although these are fairly small numbers compared to parsing, this easily shows the different performance classes for lxml and (c)ET. Where the latter do not have to care about parent pointers and tree structures, lxml has to deep traverse the appended tree. The performance difference therefore increases with the size of the tree that is moved.</p> <p>This difference is not always as visible, but applies to most parts of the API, like inserting newly created elements:</p> <pre class="literal-block"> lxe: insert_from_document (--TR T1,T2) 3.9763 msec/pass cET: insert_from_document (--TR T1,T2) 0.1459 msec/pass </pre> <p>or replacing the child slice by a newly created element:</p> <pre class="literal-block"> lxe: replace_children_element (--TC T1) 0.0749 msec/pass cET: replace_children_element (--TC T1) 0.0081 msec/pass </pre> <p>as opposed to replacing the slice with an existing element from the same document:</p> <pre class="literal-block"> lxe: replace_children (--TC T1) 0.0052 msec/pass cET: replace_children (--TC T1) 0.0036 msec/pass </pre> <p>While these numbers are too small to provide a major performance impact in practice, you should keep this difference in mind when you merge very large trees. Note that Elements have a <tt class="docutils literal">makeelement()</tt> method that allows to create an Element within the same document, thus avoiding the merge overhead when inserting it into that tree.</p> </div> <div class="section" id="deepcopy"> <h2>deepcopy</h2> <p>Deep copying a tree is fast in lxml:</p> <pre class="literal-block"> lxe: deepcopy_all (--TR T1) 3.1650 msec/pass cET: deepcopy_all (--TR T1) 53.9973 msec/pass lxe: deepcopy_all (-ATR T2) 3.7365 msec/pass cET: deepcopy_all (-ATR T2) 61.6267 msec/pass lxe: deepcopy_all (S-TR T3) 0.7913 msec/pass cET: deepcopy_all (S-TR T3) 13.6220 msec/pass </pre> <p>So, for example, if you have a database-like scenario where you parse in a large tree and then search and copy independent subtrees from it for further processing, lxml is by far the best choice here.</p> </div> <div class="section" id="tree-traversal"> <h2>Tree traversal</h2> <p>Another important area in XML processing is iteration for tree traversal. If your algorithms can benefit from step-by-step traversal of the XML tree and especially if few elements are of interest or the target element tag name is known, the <tt class="docutils literal">.iter()</tt> method is a good choice:</p> <pre class="literal-block"> lxe: iter_all (--TR T1) 1.0529 msec/pass cET: iter_all (--TR T1) 0.2635 msec/pass lxe: iter_islice (--TR T2) 0.0110 msec/pass cET: iter_islice (--TR T2) 0.0050 msec/pass lxe: iter_tag (--TR T2) 0.0079 msec/pass cET: iter_tag (--TR T2) 0.0112 msec/pass lxe: iter_tag_all (--TR T2) 0.1822 msec/pass cET: iter_tag_all (--TR T2) 0.5343 msec/pass </pre> <p>This translates directly into similar timings for <tt class="docutils literal">Element.findall()</tt>:</p> <pre class="literal-block"> lxe: findall (--TR T2) 1.7176 msec/pass cET: findall (--TR T2) 0.9973 msec/pass lxe: findall (--TR T3) 0.3967 msec/pass cET: findall (--TR T3) 0.2525 msec/pass lxe: findall_tag (--TR T2) 0.2258 msec/pass cET: findall_tag (--TR T2) 0.5770 msec/pass lxe: findall_tag (--TR T3) 0.1085 msec/pass cET: findall_tag (--TR T3) 0.1919 msec/pass </pre> <p>Note that all three libraries currently use the same Python implementation for <tt class="docutils literal">.findall()</tt>, except for their native tree iterator (<tt class="docutils literal">element.iter()</tt>). In general, lxml is very fast for iteration, but looses ground against cET when many Elements are found and need to be instantiated. So, the more selective your search is, the faster lxml will run.</p> </div> </div> <div class="section" id="xpath"> <h1>XPath</h1> <p>The following timings are based on the benchmark script <a class="reference external" href="https://github.com/lxml/lxml/blob/master/benchmark/bench_xpath.py">bench_xpath.py</a>.</p> <p>This part of lxml does not have an equivalent in ElementTree. However, lxml provides more than one way of accessing it and you should take care which part of the lxml API you use. The most straight forward way is to call the <tt class="docutils literal">xpath()</tt> method on an Element or ElementTree:</p> <pre class="literal-block"> lxe: xpath_method (--TC T1) 0.3982 msec/pass lxe: xpath_method (--TC T2) 7.8895 msec/pass lxe: xpath_method (--TC T3) 0.0477 msec/pass lxe: xpath_method (--TC T4) 0.3982 msec/pass </pre> <p>This is well suited for testing and when the XPath expressions are as diverse as the trees they are called on. However, if you have a single XPath expression that you want to apply to a larger number of different elements, the <tt class="docutils literal">XPath</tt> class is the most efficient way to do it:</p> <pre class="literal-block"> lxe: xpath_class (--TC T1) 0.0713 msec/pass lxe: xpath_class (--TC T2) 1.1325 msec/pass lxe: xpath_class (--TC T3) 0.0215 msec/pass lxe: xpath_class (--TC T4) 0.0722 msec/pass </pre> <p>Note that this still allows you to use variables in the expression, so you can parse it once and then adapt it through variables at call time. In other cases, where you have a fixed Element or ElementTree and want to run different expressions on it, you should consider the <tt class="docutils literal">XPathEvaluator</tt>:</p> <pre class="literal-block"> lxe: xpath_element (--TR T1) 0.1101 msec/pass lxe: xpath_element (--TR T2) 2.0473 msec/pass lxe: xpath_element (--TR T3) 0.0267 msec/pass lxe: xpath_element (--TR T4) 0.1087 msec/pass </pre> <p>While it looks slightly slower, creating an XPath object for each of the expressions generates a much higher overhead here:</p> <pre class="literal-block"> lxe: xpath_class_repeat (--TC T1 ) 0.3884 msec/pass lxe: xpath_class_repeat (--TC T2 ) 7.6182 msec/pass lxe: xpath_class_repeat (--TC T3 ) 0.0465 msec/pass lxe: xpath_class_repeat (--TC T4 ) 0.3877 msec/pass </pre> <p>Note that tree iteration can be substantially faster than XPath if your code short-circuits after the first couple of elements were found. The XPath engine will always return the complete result set, regardless of how much of it will actually be used.</p> <p>Here is an example where only the first matching element is being searched, a case for which XPath has syntax support as well:</p> <pre class="literal-block"> lxe: find_single (--TR T2) 0.0184 msec/pass cET: find_single (--TR T2) 0.0052 msec/pass lxe: iter_single (--TR T2) 0.0024 msec/pass cET: iter_single (--TR T2) 0.0007 msec/pass lxe: xpath_single (--TR T2) 0.0033 msec/pass </pre> <p>When looking for the first two elements out of many, the numbers explode for XPath, as restricting the result subset requires a more complex expression:</p> <pre class="literal-block"> lxe: iterfind_two (--TR T2) 0.0184 msec/pass cET: iterfind_two (--TR T2) 0.0062 msec/pass lxe: iter_two (--TR T2) 0.0029 msec/pass cET: iter_two (--TR T2) 0.0017 msec/pass lxe: xpath_two (--TR T2) 0.2768 msec/pass </pre> </div> <div class="section" id="a-longer-example"> <h1>A longer example</h1> <p>... based on lxml 1.3.</p> <p>A while ago, Uche Ogbuji posted a <a class="reference external" href="http://www.onlamp.com/pub/wlg/6291">benchmark proposal</a> that would read in a 3MB XML version of the <a class="reference external" href="http://www.ibiblio.org/bosak/xml/eg/religion.2.00.xml.zip">Old Testament</a> of the Bible and look for the word <em>begat</em> in all verses. Apparently, it is contained in 120 out of almost 24000 verses. This is easy to implement in ElementTree using <tt class="docutils literal">findall()</tt>. However, the fastest and most memory friendly way to do this is obviously <tt class="docutils literal">iterparse()</tt>, as most of the data is not of any interest.</p> <p>Now, Uche's original proposal was more or less the following:</p> <div class="syntax"><pre><span class="k">def</span> <span class="nf">bench_ET</span><span class="p">():</span> <span class="n">tree</span> <span class="o">=</span> <span class="n">ElementTree</span><span class="o">.</span><span class="n">parse</span><span class="p">(</span><span class="s">"ot.xml"</span><span class="p">)</span> <span class="n">result</span> <span class="o">=</span> <span class="p">[]</span> <span class="k">for</span> <span class="n">v</span> <span class="ow">in</span> <span class="n">tree</span><span class="o">.</span><span class="n">findall</span><span class="p">(</span><span class="s">"//v"</span><span class="p">):</span> <span class="n">text</span> <span class="o">=</span> <span class="n">v</span><span class="o">.</span><span class="n">text</span> <span class="k">if</span> <span class="s">'begat'</span> <span class="ow">in</span> <span class="n">text</span><span class="p">:</span> <span class="n">result</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">text</span><span class="p">)</span> <span class="k">return</span> <span class="nb">len</span><span class="p">(</span><span class="n">result</span><span class="p">)</span> </pre></div> <p>which takes about one second on my machine today. The faster <tt class="docutils literal">iterparse()</tt> variant looks like this:</p> <div class="syntax"><pre><span class="k">def</span> <span class="nf">bench_ET_iterparse</span><span class="p">():</span> <span class="n">result</span> <span class="o">=</span> <span class="p">[]</span> <span class="k">for</span> <span class="n">event</span><span class="p">,</span> <span class="n">v</span> <span class="ow">in</span> <span class="n">ElementTree</span><span class="o">.</span><span class="n">iterparse</span><span class="p">(</span><span class="s">"ot.xml"</span><span class="p">):</span> <span class="k">if</span> <span class="n">v</span><span class="o">.</span><span class="n">tag</span> <span class="o">==</span> <span class="s">'v'</span><span class="p">:</span> <span class="n">text</span> <span class="o">=</span> <span class="n">v</span><span class="o">.</span><span class="n">text</span> <span class="k">if</span> <span class="s">'begat'</span> <span class="ow">in</span> <span class="n">text</span><span class="p">:</span> <span class="n">result</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">text</span><span class="p">)</span> <span class="n">v</span><span class="o">.</span><span class="n">clear</span><span class="p">()</span> <span class="k">return</span> <span class="nb">len</span><span class="p">(</span><span class="n">result</span><span class="p">)</span> </pre></div> <p>The improvement is about 10%. At the time I first tried (early 2006), lxml didn't have <tt class="docutils literal">iterparse()</tt> support, but the <tt class="docutils literal">findall()</tt> variant was already faster than ElementTree. This changes immediately when you switch to cElementTree. The latter only needs 0.17 seconds to do the trick today and only some impressive 0.10 seconds when running the iterparse version. And even back then, it was quite a bit faster than what lxml could achieve.</p> <p>Since then, lxml has matured a lot and has gotten much faster. The iterparse variant now runs in 0.14 seconds, and if you remove the <tt class="docutils literal">v.clear()</tt>, it is even a little faster (which isn't the case for cElementTree).</p> <p>One of the many great tools in lxml is XPath, a swiss army knife for finding things in XML documents. It is possible to move the whole thing to a pure XPath implementation, which looks like this:</p> <div class="syntax"><pre><span class="k">def</span> <span class="nf">bench_lxml_xpath_all</span><span class="p">():</span> <span class="n">tree</span> <span class="o">=</span> <span class="n">etree</span><span class="o">.</span><span class="n">parse</span><span class="p">(</span><span class="s">"ot.xml"</span><span class="p">)</span> <span class="n">result</span> <span class="o">=</span> <span class="n">tree</span><span class="o">.</span><span class="n">xpath</span><span class="p">(</span><span class="s">"//v[contains(., 'begat')]/text()"</span><span class="p">)</span> <span class="k">return</span> <span class="nb">len</span><span class="p">(</span><span class="n">result</span><span class="p">)</span> </pre></div> <p>This runs in about 0.13 seconds and is about the shortest possible implementation (in lines of Python code) that I could come up with. Now, this is already a rather complex XPath expression compared to the simple "//v" ElementPath expression we started with. Since this is also valid XPath, let's try this instead:</p> <div class="syntax"><pre><span class="k">def</span> <span class="nf">bench_lxml_xpath</span><span class="p">():</span> <span class="n">tree</span> <span class="o">=</span> <span class="n">etree</span><span class="o">.</span><span class="n">parse</span><span class="p">(</span><span class="s">"ot.xml"</span><span class="p">)</span> <span class="n">result</span> <span class="o">=</span> <span class="p">[]</span> <span class="k">for</span> <span class="n">v</span> <span class="ow">in</span> <span class="n">tree</span><span class="o">.</span><span class="n">xpath</span><span class="p">(</span><span class="s">"//v"</span><span class="p">):</span> <span class="n">text</span> <span class="o">=</span> <span class="n">v</span><span class="o">.</span><span class="n">text</span> <span class="k">if</span> <span class="s">'begat'</span> <span class="ow">in</span> <span class="n">text</span><span class="p">:</span> <span class="n">result</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">text</span><span class="p">)</span> <span class="k">return</span> <span class="nb">len</span><span class="p">(</span><span class="n">result</span><span class="p">)</span> </pre></div> <p>This gets us down to 0.12 seconds, thus showing that a generic XPath evaluation engine cannot always compete with a simpler, tailored solution. However, since this is not much different from the original findall variant, we can remove the complexity of the XPath call completely and just go with what we had in the beginning. Under lxml, this runs in the same 0.12 seconds.</p> <p>But there is one thing left to try. We can replace the simple ElementPath expression with a native tree iterator:</p> <div class="syntax"><pre><span class="k">def</span> <span class="nf">bench_lxml_getiterator</span><span class="p">():</span> <span class="n">tree</span> <span class="o">=</span> <span class="n">etree</span><span class="o">.</span><span class="n">parse</span><span class="p">(</span><span class="s">"ot.xml"</span><span class="p">)</span> <span class="n">result</span> <span class="o">=</span> <span class="p">[]</span> <span class="k">for</span> <span class="n">v</span> <span class="ow">in</span> <span class="n">tree</span><span class="o">.</span><span class="n">getiterator</span><span class="p">(</span><span class="s">"v"</span><span class="p">):</span> <span class="n">text</span> <span class="o">=</span> <span class="n">v</span><span class="o">.</span><span class="n">text</span> <span class="k">if</span> <span class="s">'begat'</span> <span class="ow">in</span> <span class="n">text</span><span class="p">:</span> <span class="n">result</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">text</span><span class="p">)</span> <span class="k">return</span> <span class="nb">len</span><span class="p">(</span><span class="n">result</span><span class="p">)</span> </pre></div> <p>This implements the same thing, just without the overhead of parsing and evaluating a path expression. And this makes it another bit faster, down to 0.11 seconds. For comparison, cElementTree runs this version in 0.17 seconds.</p> <p>So, what have we learned?</p> <ul class="simple"> <li>Python code is not slow. The pure XPath solution was not even as fast as the first shot Python implementation. In general, a few more lines in Python make things more readable, which is much more important than the last 5% of performance.</li> <li>It's important to know the available options - and it's worth starting with the most simple one. In this case, a programmer would then probably have started with <tt class="docutils literal"><span class="pre">getiterator("v")</span></tt> or <tt class="docutils literal">iterparse()</tt>. Either of them would already have been the most efficient, depending on which library is used.</li> <li>It's important to know your tool. lxml and cElementTree are both very fast libraries, but they do not have the same performance characteristics. The fastest solution in one library can be comparatively slow in the other. If you optimise, optimise for the specific target platform.</li> <li>It's not always worth optimising. After all that hassle we got from 0.12 seconds for the initial implementation to 0.11 seconds. Switching over to cElementTree and writing an <tt class="docutils literal">iterparse()</tt> based version would have given us 0.10 seconds - not a big difference for 3MB of XML.</li> <li>Take care what operation is really dominating in your use case. If we split up the operations, we can see that lxml is slightly slower than cElementTree on <tt class="docutils literal">parse()</tt> (both about 0.06 seconds), but more visibly slower on <tt class="docutils literal">iterparse()</tt>: 0.07 versus 0.10 seconds. However, tree iteration in lxml is increadibly fast, so it can be better to parse the whole tree and then iterate over it rather than using <tt class="docutils literal">iterparse()</tt> to do both in one step. Or, you can just wait for the lxml developers to optimise iterparse in one of the next releases...</li> </ul> </div> <div class="section" id="lxml-objectify"> <h1>lxml.objectify</h1> <p>The following timings are based on the benchmark script <a class="reference external" href="https://github.com/lxml/lxml/blob/master/benchmark/bench_objectify.py">bench_objectify.py</a>.</p> <p>Objectify is a data-binding API for XML based on lxml.etree, that was added in version 1.1. It uses standard Python attribute access to traverse the XML tree. It also features ObjectPath, a fast path language based on the same meme.</p> <p>Just like lxml.etree, lxml.objectify creates Python representations of elements on the fly. To save memory, the normal Python garbage collection mechanisms will discard them when their last reference is gone. In cases where deeply nested elements are frequently accessed through the objectify API, the create-discard cycles can become a bottleneck, as elements have to be instantiated over and over again.</p> <div class="section" id="objectpath"> <h2>ObjectPath</h2> <p>ObjectPath can be used to speed up the access to elements that are deep in the tree. It avoids step-by-step Python element instantiations along the path, which can substantially improve the access time:</p> <pre class="literal-block"> lxe: attribute (--TR T1) 4.1828 msec/pass lxe: attribute (--TR T2) 17.3802 msec/pass lxe: attribute (--TR T4) 3.8657 msec/pass lxe: objectpath (--TR T1) 0.9289 msec/pass lxe: objectpath (--TR T2) 13.3109 msec/pass lxe: objectpath (--TR T4) 0.9289 msec/pass lxe: attributes_deep (--TR T1) 6.2900 msec/pass lxe: attributes_deep (--TR T2) 20.4713 msec/pass lxe: attributes_deep (--TR T4) 6.1679 msec/pass lxe: objectpath_deep (--TR T1) 1.3049 msec/pass lxe: objectpath_deep (--TR T2) 14.0815 msec/pass lxe: objectpath_deep (--TR T4) 1.3051 msec/pass </pre> <p>Note, however, that parsing ObjectPath expressions is not for free either, so this is most effective for frequently accessing the same element.</p> </div> <div class="section" id="caching-elements"> <h2>Caching Elements</h2> <p>A way to improve the normal attribute access time is static instantiation of the Python objects, thus trading memory for speed. Just create a cache dictionary and run:</p> <div class="syntax"><pre><span class="n">cache</span><span class="p">[</span><span class="n">root</span><span class="p">]</span> <span class="o">=</span> <span class="nb">list</span><span class="p">(</span><span class="n">root</span><span class="o">.</span><span class="n">iter</span><span class="p">())</span> </pre></div> <p>after parsing and:</p> <div class="syntax"><pre><span class="k">del</span> <span class="n">cache</span><span class="p">[</span><span class="n">root</span><span class="p">]</span> </pre></div> <p>when you are done with the tree. This will keep the Python element representations of all elements alive and thus avoid the overhead of repeated Python object creation. You can also consider using filters or generator expressions to be more selective. By choosing the right trees (or even subtrees and elements) to cache, you can trade memory usage against access speed:</p> <pre class="literal-block"> lxe: attribute_cached (--TR T1) 3.1357 msec/pass lxe: attribute_cached (--TR T2) 15.8911 msec/pass lxe: attribute_cached (--TR T4) 2.9194 msec/pass lxe: attributes_deep_cached (--TR T1) 3.8984 msec/pass lxe: attributes_deep_cached (--TR T2) 16.8300 msec/pass lxe: attributes_deep_cached (--TR T4) 3.6936 msec/pass lxe: objectpath_deep_cached (--TR T1) 0.7496 msec/pass lxe: objectpath_deep_cached (--TR T2) 12.3763 msec/pass lxe: objectpath_deep_cached (--TR T4) 0.7427 msec/pass </pre> <p>Things to note: you cannot currently use <tt class="docutils literal">weakref.WeakKeyDictionary</tt> objects for this as lxml's element objects do not support weak references (which are costly in terms of memory). Also note that new element objects that you add to these trees will not turn up in the cache automatically and will therefore still be garbage collected when all their Python references are gone, so this is most effective for largely immutable trees. You should consider using a set instead of a list in this case and add new elements by hand.</p> </div> <div class="section" id="further-optimisations"> <h2>Further optimisations</h2> <p>Here are some more things to try if optimisation is required:</p> <ul class="simple"> <li>A lot of time is usually spent in tree traversal to find the addressed elements in the tree. If you often work in subtrees, do what you would also do with deep Python objects: assign the parent of the subtree to a variable or pass it into functions instead of starting at the root. This allows accessing its descendents more directly.</li> <li>Try assigning data values directly to attributes instead of passing them through DataElement.</li> <li>If you use custom data types that are costly to parse, try running <tt class="docutils literal">objectify.annotate()</tt> over read-only trees to speed up the attribute type inference on read access.</li> </ul> <p>Note that none of these measures is guaranteed to speed up your application. As usual, you should prefer readable code over premature optimisations and profile your expected use cases before bothering to apply optimisations at random.</p> </div> </div> </div> <div class="footer"> <hr class="footer" /> Generated on: 2013-11-07. </div> </body> </html>