print text inside .. and ..

Question

I need to list the text inside these two elements in the many pom.xml in a directory tree. The files may contain the element at other places too, I am only looking for the content of these two.

Ideally, I am looking for a tool that outputs in the format <file-name>:<line-no>:<path>:<text>, e.g.

parent/pom.xml:12:/project/groupId:com.acme.project
features/persist/pom.xml:14:/project/parent/groupId:com.acme.project
features/persist/pom.xml:32:/project/groupId:com.acme.project.persist

For the following input files:

**parent/pom/xml**
<project>
 ...
  <groupId>
  com.acme.project <!--LINE 12 --> 
  </groupId>
...
</project>

**feature/persist/pom.xml**
<project>
  <parent>
    <groupId>
    com.acme.project <!--LINE 14 --> 
    </groupId>
  </parent>
  ...
  <groupId>
  com.acme.project
  </groupId>
  <dependencyManagement>
    <dependencies>
      <dependency>
        <groupId>
        a.b.c.d <!-- this is not listed in output -->
        </groupId>
      </dependency>
    </dependencies>
  </dependencyManagement>
</project>

Note that other paths such as dependencyManagement/dependencies/dependency/groupId are not included.

Searching here on SO I came across xmllint --xpath and xmlstarlet but I don't know enough about xpath to figure this out myself.

Output was made up :) But I have made up the corresponding input and added it — Miserable Variable, Apr 18 '13 at 19:45
Just realized you need the line number and path to the element. There is no way to get the line number by standard XPath/XQuery, not even in version 3.0. At least [Saxon has some propietary support for this](http://saxon.sourceforge.net/saxon6.5.3/extensions.html#linenumber). To the path: there is `fn:path()` for this in XPath/XQuery 3.0, so you will need a more capable and up to date processor. Neither [xmllint] nor [xmlstartlet] support more than XPath 1.0, so you will need another tool. — Jens Erat, Apr 18 '13 at 21:34

score 0 · Answer 1 · answered Apr 21 '13 at 13:20

Try this XPath 2.0 compatible query without line numbers, which checks all query paths for matching <groupID/> elements and prints their document name, constructs a (not necessarily unique) path and adds the contents of the element.

(//project/parent | //project)/groupId/string-join(
  (
    base-uri(),
    string-join(('', .//ancestor-or-self::*/name()), '/'),
    data(.)
  ), ':')

You could run it using a BaseX collection for example (like I did for testing), which contains all the XML files you want to query.

Run this command to create the collection: CREATE DB xmldocs /path/to/xml-files
Query the database using above XPath

There are different ways to run the query, have a look at the Standalone Mode manual.

The query should also run in other XPath 2.0 compatible engines like saxon (which also would support line numbers, see my comment above).

score 0 · Answer 2 · answered May 08 '13 at 18:56

0

I ended up using the cygwin build of xml2:

xml2 <  pom.xml | grep -e "/project/parent/groupId" -e "/project/groupId"
/project/parent/groupId=...
/project/groupId=....

answered May 08 '13 at 18:56

Miserable Variable

28,432
15
72
133

print text inside .. and ..

2 Answers2