XML Parsing with Xpath(?) for certain value

Question

I'm needing to parse a particular set of XML files for a value.

In this instance, I need to find every access id for <group>THISGROUP</group> with <permission>WRITE</permission>

In the example below, I would need VX-3422867 returned. Also, I need to be able to accomplish this within a bash script.

<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<AccessControlListDocument>
    <access id="VX-3422867">
        <loc>http://localhost:8080/API/item/VX-410070/access/VX-3422867</loc>
        <grantor>admin</grantor>
        <recursive>true</recursive>
        <permission>WRITE</permission>
        <group>THISGROUP</group>
    </access>
        <access id="VX-3422866">
        <loc>http://localhost:8080/API/item/VX-410070/access/VX-3422866</loc>
        <grantor>admin</grantor>
        <recursive>true</recursive>
        <permission>NONE</permission>
        <operation>
            <shape>
                <tag>original</tag>
            </shape>
        </operation>
        <group>THISGROUP</group>
    </access>
    <access id="VX-3422865">
        <loc>http://localhost:8080/API/item/VX-410070/access/VX-3422865</loc>
        <grantor>admin</grantor>
        <recursive>true</recursive>
        <permission>WRITE</permission>
        <group>TWO</group>
    </access>
    <access id="VX-3422869">
        <loc>http://localhost:8080/API/item/VX-410070/access/VX-3422869</loc>
        <grantor>admin</grantor>
        <recursive>true</recursive>
        <permission>ALL</permission>
        <group>THREE</group>
    </access>
    <access id="VX-3422868">
        <loc>http://localhost:8080/API/item/VX-410070/access/VX-3422868</loc>
        <grantor>admin</grantor>
        <recursive>true</recursive>
        <permission>ALL</permission>
        <group>FOUR</group>
    </access>
    <access id="VX-975588">
        <loc>http://localhost:8080/API/item/VX-410070/access/VX-975588</loc>
        <recursive>true</recursive>
        <permission>OWNER</permission>
        <user>user-one</user>
    </access>
</AccessControlListDocument>

You have a lot of reading to do, methinks. You should start perhaps [here](http://stackoverflow.com/questions/4984689/bash-xhtml-parsing-using-xpath). — Shawn Mehan, Sep 02 '15 at 01:07
BTW, "this is my input, please write a program for me that gives me the output I want" is a subcategory of "please do my work for me", which is not particularly welcome here -- hence the downvotes. Making an effort to at least show what you've tried or considered is helpful. — Charles Duffy, Sep 02 '15 at 02:38
My apologies for not explaining everything I had tried. I'm sure you understand being in a pinch and needing help. — CEC, Sep 03 '15 at 13:56

score 2 · Answer 1 · answered Sep 02 '15 at 02:36

2

To print the IDs for all access entries with groups THISGROUP and permission WRITE:

xmlstarlet sel \
  -t -m '//access[group="THISGROUP"][permission="WRITE"]' \
  -v @id -n <in.xml

answered Sep 02 '15 at 02:36

Charles Duffy

280,126
43
390
441

score 2 · Answer 2 · answered Sep 02 '15 at 02:39

You can select the @id attributes of the access elements who's group element value is equal to 'THISGROUP' and permission element value is equal to 'WRITE' with the following XPath:

/AccessControlListDocument/access[group = 'THISGROUP' and permission='WRITE']/@id

As for executing in a bash script, you could look at using xmlstarlet:

xmlstarlet sel -t -v "/AccessControlListDocument/access[group = 'THISGROUP' and permission='WRITE']/@id" myFile.xml

Or with a variety of other options (xmllint, saxon-lint, saxon, python, xidel) as suggested in answers to How to execute XPath one-liners from shell?

Thank you, I'm used to using Xidel. Turns out I was close to the syntax I needed, but not quite there. — CEC, Sep 03 '15 at 13:56

score 0 · Answer 3 · answered Sep 02 '15 at 03:09

0

whith xpath:

If you wanna only de first id atribute:

   <xsl:value-of select="/AccessControlListDocument/access[child::group='THISGROUP' and child::permission='WRITE']/@id"/>

If you wanna every id atribute whit that condition:

<xsl:for-each select="/AccessControlListDocument/access[child::group='THISGROUP' and child::permission='WRITE']">
<xsl:value-of select="./@id"/>
</xsl:for-each>

This means get the "id" atribute from "access" node child from "AccessControlListDocument" parent, where "access" node have children nodes "group" with value 'THISGROUP' and node "permission" with value 'WRITE'.

answered Sep 02 '15 at 03:09

Mauro

93
8

This is actually a very good approach, as people are more likely to have xsltproc than they are to have xmlstarlet or an xmllint recent enough to have `--xpath`. You should perhaps show how to run it from bash, though. – Charles Duffy Sep 02 '15 at 03:10
I'd also suggest describing it as "With XSLT" rather than "with XPath". – Charles Duffy Sep 02 '15 at 03:13

XML Parsing with Xpath(?) for certain value

3 Answers3