Liking cljdoc? Tell your friends :D

xml-picker-seq

This wrapper tries to make it easy to pull out interesting bits of large XML files in a space-efficient way.

Almost every XML file I ever look at seems to be large (>1GB) and oriented around records (MARC records, Solr documents, etc.). Generally all I want to do is:

  • Read the next record
  • Pull out a few of its fields
  • Throw the rest away, rinse and repeat

so that's what this does, using XOM and a queue behind the scenes.

Example usage

Parsing a Solr response

<response>
  <lst name="responseHeader">
    <int name="status">0</int>
    <int name="QTime">0</int>
  </lst>
  <result name="response" numFound="170" start="0">
    <doc>
      <arr name="author"><str>Hardie &amp; Gorman Pty. Ltd</str></arr>
      <arr name="callnumber"><str>MAP FOLDER 176, LFSP 2757</str></arr>
      <arr name="title"><str>Important clearance sale of valuable city, suburban &amp; country properties, at upset prices : including investment, residential and vacant properties, building, farming, agricultural, and grazing lands ... Auction sale, Wednesday, 8th December, 1897 / Hardie &amp; Gorman, Auctioneers</str></arr>
    </doc>
    <doc>
      <arr name="author"><str>Hamburg, Harold G</str></arr>
      <arr name="callnumber"><str>PAM BOX 429</str></arr>
      <arr name="title"><str>The legislation of Richard III</str></arr>
    </doc>
    ...
  </result>
</response>

to extract a list of authors:

(with-open [rdr (clojure.java.io/reader (java.net.URL. "http://my.host/solr/select?q=my+query"))]
  (doall (apply concat
                (xml-picker-seq.core/xml-picker-seq
                 rdr
                "doc"
                (xml-picker-seq.core/xpath-query "arr[@name='author']/str")))))

This:

  • Walks through the document parsing (and loading into memory) one full "doc" node at a time

  • Uses XPath to pull out all of the str values for "author"

  • Gets the string values for each author node and returns them

Parsing MARCXML records

Slightly more fiddly because of the namespaces, but much the same:

(with-open [rdr (clojure.java.io/reader "http://www.loc.gov/standards/marcxml/xml/collection.xml")]
  (let [context (nu.xom.XPathContext. "marc" "http://www.loc.gov/MARC21/slim")
        titles (xml-picker-seq.core/xml-picker-seq
                rdr "record"
                (xml-picker-seq.core/xpath-query "//marc:datafield[@tag = '245']/marc:subfield[@code = 'a']"
                                                 :context context :final-fn first))]
    (doseq [title titles]
      (do-something-with title))))

Parsing KML files

(ns kml-parser.core
  (:use [xml-picker-seq.core :only [xml-picker-seq xpath-query]]
        [clojure.java.io :only [reader]])
  (:require [clojure.string :as string]))

(defn read-kml [filename]
  (with-open [rdr (reader filename)]
    (doall (xml-picker-seq rdr "Placemark"
                           (juxt (xpath-query "*[local-name()='Point']/*[local-name()='coordinates']" :final-fn first)
                                 (xpath-query "*[local-name()='description']" :final-fn first
                                              :extract-fn #(-> % .getValue (string/split #"\n"))))))))

TODO

  • Replace clojure.contrib.seq-utils/fill-queue, contrib is depreciated.

License

Copyright (C) 2009 Mark Triggs

Distributed under the Eclipse Public License. See the file "COPYING".

Can you improve this documentation? These fine people already did:
David Barker & Mark Triggs
Edit on GitHub

cljdoc is a website building & hosting documentation for Clojure/Script libraries

× close