Scripting: what is the easiest to extract a value in a tag of a XML file?

Question

I want to read a pom.xml ('Project Object Model' of Maven) and extract the version information. Here is an example:

<?xml version="1.0" encoding="UTF-8"?><project 
xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/maven-v4_0_0.xsd">

    <modelVersion>4.0.0</modelVersion>
    <groupId>com.mycompany</groupId>
    <artifactId>project-parent</artifactId>
    <name>project-parent</name>
    <version>1.0.74-SNAPSHOT</version>
    <dependencies>
        <dependency>
        <groupId>com.sybase.jconnect</groupId>
        <artifactId>jconnect</artifactId>
        <version>6.05-26023</version>
    </dependency>
    <dependency>
        <groupId>joda-time</groupId>
        <artifactId>joda-time</artifactId>
        <version>1.5.2</version>
    </dependency>
    <dependency>
        <groupId>com.sun.jdmk</groupId>
        <artifactId>jmxtools</artifactId>
        <version>1.2.1</version>
    </dependency>
    <dependency>
        <groupId>org.easymock</groupId>
        <artifactId>easymock</artifactId>
        <version>2.4</version>
    </dependency>       
</dependencies>
</project>

How can I extract the version '1.0.74-SNAPSHOT' from above?

Would love to be able to do so using simple bash scripting sed or awk. Otherwise a simple python is preferred.

EDIT

1. Constraint

   The linux box is in a corporate environment so I can only use tools that are already installed (not that I cannot request utility such as xml2, but I have to go through a lot of red-tape). Some of the solutions are very good (learn a few new tricks already), but they may not be applicable due to the restricted environment

2. updated xml listing

   I added the dependencies tag to the original listing. This will show some hacky solution may not work in this case

3. Distro

   The distro I am using is RHEL4

Solution 1

xml2 can convert xml to/from line-oriented format:

xml2 < pom.xml  | grep /project/version= | sed 's/.*=//'

Solution 2

Other way: xmlgrep and XPath:

xmlgrep --text_only '/project/version' pom.xml

Disadvantage: slow

Solution 3

Using python

$ python -c 'from xml.etree.ElementTree import ElementTree; print ElementTree(file="pom.xml").findtext("{http://maven.apache.org/POM/4.0.0}version")'
1.0.74-SNAPSHOT

Using xmlstarlet

$ xml sel -N x="http://maven.apache.org/POM/4.0.0" -t -m 'x:project/x:version' -v . pom.xml
1.0.74-SNAPSHOT

Using xmllint

$ echo -e 'setns x=http://maven.apache.org/POM/4.0.0\ncat /x:project/x:version/text()' | xmllint --shell pom.xml | grep -v /
1.0.74-SNAPSHOT

Solution 4

Clojure way. Requires only jvm with special jar file:

java -cp clojure.jar clojure.main -e "(use 'clojure.xml) (->> (java.io.File. \"pom.xml\") (clojure.xml/parse) (:content) (filter #(= (:tag %) :version)) (first) (:content) (first) (println))"

Scala way:

java -Xbootclasspath/a:scala-library.jar -cp scala-compiler.jar scala.tools.nsc.MainGenericRunner -e 'import scala.xml._; println((XML.load(new java.io.FileInputStream("pom.xml")) match { case <project>{children @ _*}</project> => for (i <- children if (i  match { case <version>{children @ _*}</version> => true; case _ => false;  }))  yield i })(0) match { case <version>{Text(x)}</version> => x })'

Groovy way:

java -classpath groovy-all.jar groovy.ui.GroovyMain -e 'println (new XmlParser().parse(new File("pom.xml")).value().findAll({ it.name().getLocalPart()=="version" }).first().value().first())'

Solution 5

Here's an alternative in Perl

$ perl -MXML::Simple -e'print XMLin("pom.xml")->{version}."\n"'
1.0.74-SNAPSHOT

It works with the revised/extended example in the questions which has multiple "version" elements at different depths.

View more solutions

Comments

[bbaja42]

Is this stackoverflow.com/questions/29004/… sufficient?

[Anthony Kong]

Not really. There are a lot of version tag in the xml (e.g. under dependencies tag). I only want '/project/version'

[Vi.]

Which xml-related tools and libraries are available? Are jvm-based soltuions OK?

[Anthony Kong]

So far I can tell xml2, xmlgrep and perl XML module are not present. Most unix command-line utilities are present. The distro is Redhat EL 4.

[JStrahl]

(I couldn't add a comment so I have to reply as an answer, overkill somewhat) Some great answers can be found here..... stackoverflow.com/questions/2735548/…

[Ciro Santilli Путлер Капут 六四事]

stackoverflow.com/questions/893585/how-to-parse-xml-in-bash

[Vi.]

Slow, (although faster than xmlgrep)

[Anthony Kong]

Thanks for the suggestion, but unfortunately it will not return what I want. Please see the updated pom model.

[Vi.]

Returns "1.0.74-SNAPSHOT". Note that I changed the script after reading about multiple <version> things.

[Vi.]

Note: this solution is provided "just for fun" and is not intended to be used in actual product. Better use xml2/xmlgrep/XML::Simple solution.

[Anthony Kong]

Thanks! even though it is 'just for fun' but it is probably the 'most suitable' solution by far because it has minimum number of dependencies: It only requires perl ;-)

[Vi.]

What about doing it from Java? Using pom files implies having JVM installed.

[Anthony Kong]

The background is that I am building a SIT (system integration test) script around the existing maven process. Part of it requires knowing the version of the maven project. I really want to keep it simple and scripting is the way to go.

[Anthony Kong]

This is awesome! Great idea!

[David H]

If xsltproc is on your system, and it probably is as libxslt is on RHEL4, then you can use it and the above stylesheet to output the tag, i.e. xsltproc x.xsl prom.xsl.

[kev]

cat (//x:version)[1]/text() when using xmllint also works!

[Vi.]

Relies on absence of parameters in elements and that extra <version>s can be only inside dependencies.

[Simon Sheehan]

What exactly does this script do?

[Samus_]

it loads the XML as a DOM structure using Python's minidom implementation: docs.python.org/library/xml.dom.minidom.html the idea is to grab the <project> tag that is unique and then iterate over its child nodes (direct childs only) to find the tag <version> that we're looking for and not other tags with the same name in other places.

[fixer1234]

Can you expand your answer to explain this? Thanks.

[GAD3R]

command updated to xml_grep

[Charlweed]

Powershell is now open source and runs on Linux and other platforms. We use it for building in preference to bash, cygwin and ming64.

[SMerrill8]

This does appear to work, but beware: What it does is set the field separator (FS) to the set of characters < and >; then it finds all lines with the word "packaging" in them and give you the third field.

[user5249203]

xml2 can be found at github.com/clone/xml2 - it's original website etc have disappeared.