52
  • I don't want to create an XML file
  • I need to use the --shell with cat to create filters

  • I don't have xpath in my version of xmllint, using libxml2-2.7.6-14.el6.x86_64

    xml|xmllint --shell - <<< $(echo 'cat /')

    -:1: parser error : Start tag expected, '<' not found

EDIT: clustat -x outputs an XML file and I want to parse out the active node. I don't think there is a way to do it without xpath so I created a temp xml file.

/usr/sbin/clustat -x > /tmp/clustat.xml
ACTIVENODE=$(xmllint --shell /tmp/clustat.xml <<< `echo 'cat //group/@owner'`|grep -v "^/ >"|cut -d= -f2|tr -d \")
Ura
  • 2,173
  • 3
  • 24
  • 41

1 Answers1

70

I had a similar issue where I had to unzip an XML file then feed it to xmllint. The key is the "-" option which tells xmllint to read from stdin.

For example:

$ bzip2 -dc dierehabilitati00delagoog_meta.xml.bz2 | xmllint --format

would fail giving the "usage" for xmllint. Adding "-" worked:

$ bzip2 -dc dierehabilitati00delagoog_meta.xml.bz2 | xmllint --format -
<?xml version="1.0"?>
<metadata>
  <title>Die Rehabilitation im Strafrecht</title>
  <creator>Ernst Delaquis</creator>
  <mediatype>texts</mediatype>
  <collection>americana</collection>
</metadata>

Hope this helps.

aiGuru
  • 1,110
  • 1
  • 7
  • 15
  • 1
    You are right, this will work from stdin but can you also create filters? – Ura Dec 13 '13 at 20:00
  • I'm not familiar with clustat, but I used sed to pull the values for a tag. Ex: bzip2 -dc dierehabilitati00delagoog_meta.xml.bz2 | xmllint --format - | sed -n 's|<[creator]*>\(.*\)<[creator/]*>|\1|p' would return "Ernst Delaquis" – aiGuru Dec 13 '13 at 21:06