clj-xpath is a library that makes it easier to work with XPath from Clojure.
Documentation is available here
Simplified XPath Library for Clojure. XML Parsers and an XPath implementation now comes with Java 6, though using the api directly can be verbose and confusing. This library provides a thin layer around basic parsing and XPath interaction for common use cases. I have personally found the ability to interactively tweak my xpath expressions to be a great productivity boost – even using this library only for that has helped me in my learning of and using xpath. I hope you find it useful and would love to hear your feedback and suggestions.
The main functions in the library are $x
and those named with a prefix of $x:
(eg: $x:text
). The rationale for choosing $x
as a name was based on the FireBug xpath function and it being a short and uncommon name. These xpath functions all take the xpath expression to be executed and an XML document. They attempt to be flexible with respect to the form of the XML document may represent. If it is a string it is treated as XML, if a byte array it is used directly, if already a Document or Node (from org.w3c.dom) they are used as-is.
There are four forms of most of the core functions, each with a different suffix borrowed from regular expression syntax: none, * + and ?. For example, $x:tag
has the following four implementations:
($x:tag "//books")
: ‘1 and only 1’, returns the single node found, throwing an exception if none or more than 1 are found.($x:tag? "//books")
: ‘0 or 1’, returns the single node found or nil, throwing an exception if more than 1 are found.($x:tag* "//book")
: ‘0 or more’, returns a sequence of the nodes found (which may be empty)($x:tag+ "//book")
: ‘1 or more’returns a sequence of the nodes found, throwing an exception if none are found
If you are interested in the entire node found by the XPath expressions and not just in particular aspects the node (tag, attributes, text content), $x
function returns a map containing the XML tag (as a symbol), dom Node, the text (as a string), and a map of the attributes where the keys have been converted into keywords and the values remain Strings.
(ns example
(use [clj-xpath.core :only [$x $x:tag $x:text $x:attrs $x:attrs* $x:node]]))
(def *some-xml*
"<?xml version=\"1.0\" encoding=\"UTF-8\"?>
<books>
<book title=\"Some Guide To XML\">
<author>
<name>P.T. Xarnum</name>
<email>[email protected]</email>
</author>
<description>
Simply the most comprehensive XML Book on the market today.
</description>
</book>
<book title=\"Some Guide To Functional Programming\">
<author>
<name>S. Hawking</name>
<email>[email protected]</email>
</author>
<description>
This book is too smart for you, try 'Head first Quantum Mechanics for Dummies' instead.
</description>
</book>
</books>")
;; get the top level tag:
(prn ($x:tag "/*" *some-xml*))
;; :books
;; find all :book nodes, pull the title from the attributes:
(prn (map #(-> % :attrs :title) ($x "//book" *some-xml*)))
;; ("Some Guide To XML" "Some Guide To Functional Programming")
;; same result using the $x:attrs* function:
(prn ($x:attrs* "//book" *some-xml* :title))
;; ("Some Guide To XML" "Some Guide To Functional Programming")
;; first select the :book element who's title has 'XML' in it
;; from that node, get and print the author's name (text content):
(prn ($x:text "./author/name"
($x:node "//book[contains(@title,'XML')]" *some-xml*)))
;; "P.T. Xarnum"
The $x
and related functions support Strings, and in many cases, other convenient types for these arguments. In all cases where it expects an XML Document it can be given a String, a byte array or a Document. Where an xpath expression is expected it will take either a String or a pre-compiled XPathExpression. The act of parsing an XML document or compiling an xpath expression is an expensive activity. With this flexibility, clj-xpath supports the convenience of in-line usage (with String data), as well as pre-parsed and pre-compiled instances for better performance.
(let [expr (xp:compile "/*")
doc (xml->doc "<authors><author><name>P.T. Xarnum</name></author></authors>")]
($x:tag expr doc))
This function takes xml that is of one of the following types and returns a Document: String, byte array or org.w3c.dom.Document. In cases of repeated usage of the document (eg: executing multiple xpath expressions against the same document) this will improve performance.
Pre-compiles the xpath expression. In cases of repeated execution of the xpath expression this will improve performance.
Validation now off by default. Validation is controlled by optional parameters passed to xml-bytes->dom
, or by overriding the atom *validation*
to false:
(ns your.namespace
(:use clj-xpath.core))
(binding [*validation* false]
($x:text "/this" "<this>foo</this>"))
To use the xpath library with an XML document that utilizes XML namespaces, you can make use of the `with-namespace-context` macro providing a map of aliases to the xmlns URL:
(def xml (slurp "fixtures/namespace1.xml"))
(with-namespace-context {"atom" "http://www.w3.org/2005/Atom"}
($x:text "//atom:title" xml-doc))
;; => BookingCollection
There is also a utility function that can pull the namespace declarations from the root node of your XML document:
(def xml (slurp "fixtures/namespace1.xml"))
(with-namespace-context (xmlnsmap-from-root-node xml-doc)
($x:text "//atom:title" xml-doc))
;; => BookingCollection
These two examples assume the following XML document:
<atom:feed xml:base="http://nplhost:8042/sap/opu/sdata/IWFND/RMTSAMPLEFLIGHT/"
xmlns:atom="http://www.w3.org/2005/Atom"
xmlns:d="http://schemas.microsoft.com/ado/2007/08/dataservices"
xmlns:m="http://schemas.microsoft.com/ado/2007/08/dataservices/metadata"
xmlns:sap="http://www.sap.com/Protocols/SAPData">
<atom:title>BookingCollection</atom:title>
<atom:updated>2012-03-19T20:27:30Z</atom:updated>
<atom:entry>
<atom:author/>
</atom:entry>
<atom:entry>
<atom:author/>
<atom:content type="application/xml"/>
</atom:entry>
</atom:feed>
- Compatibility with Clojure 1.2, 1.3, 1.4, 1.5 and 1.6-SNAPSHOT
- Support leiningen 2
- create profiles for clojure 1.2 through 1.6
- resolve reflection warnings: NB: two remain for clojure 1.3
Changed project group from org.clojars.kyleburton to com.github.kyleburton.
:children
lazy seq of a Node’s children added by mtnygard- idiomatic use of next
- Kyle Burton <[email protected]>
- Trotter Cashion <[email protected]>
- Michael Nygard <[email protected]>