Parsing XML from a shell script

Question

I have a datafile like this:

<Key name="com.ahsay.afc.cpf.UserGroup" content="" allowMultiple="Y"> <Value name="rsv-id" inheritParentAttribute="Y" type="string" data="1328200856753" /> <Value name="rsv-group-name" inheritParentAttribute="Y" type="string" data="group 1" /> <Value name="rsv-user-type" inheritParentAttribute="Y" type="string" data="backup-user" /> <Value name="rsv-owner" inheritParentAttribute="Y" type="string" data="" /> <Key name="com.ahsay.afc.cpf.User" content="" allowMultiple="Y"> <Value name="rsv-id" inheritParentAttribute="Y" type="string" data="13279083887401" /> <Value name="rsv-login-name" inheritParentAttribute="Y" type="string" data="name1" /> </Key> <Key name="com.ahsay.afc.cpf.User" content="" allowMultiple="Y"> <Value name="rsv-id" inheritParentAttribute="Y" type="string" data="13279083887401" /> <Value name="rsv-login-name" inheritParentAttribute="Y" type="string" data="name2" /> </Key> </Key> <Key name="com.ahsay.afc.cpf.UserGroup" content="" allowMultiple="Y"> <Value name="rsv-id" inheritParentAttribute="Y" type="string" data="1328200856753" /> <Value name="rsv-group-name" inheritParentAttribute="Y" type="string" data="group 2" /> <Value name="rsv-user-type" inheritParentAttribute="Y" type="string" data="backup-user" /> <Value name="rsv-owner" inheritParentAttribute="Y" type="string" data="" /> <Key name="com.ahsay.afc.cpf.User" content="" allowMultiple="Y"> <Value name="rsv-id" inheritParentAttribute="Y" type="string" data="13279083887401" /> <Value name="rsv-login-name" inheritParentAttribute="Y" type="string" data="name3" /> </Key> <Key name="com.ahsay.afc.cpf.User" content="" allowMultiple="Y"> <Value name="rsv-id" inheritParentAttribute="Y" type="string" data="13279083887401" /> <Value name="rsv-login-name" inheritParentAttribute="Y" type="string" data="name4" /> </Key> </Key> 

I know the login name of the record I want, and I need to match it to a group. Let's say I want to know what group name3 is (the answer is group 3). Currently I can get the name of the group out of the file with:

perl -ne 'print "$_\n" foreach /name="rsv-group-name".*\ data="([^"]*)"/g;' 

but I have no idea how to match it with a user. How can I do that in a script?

Answer 1

The group is group 2, not group 3:

xmlstarlet sel -t \ -v '//Key[Key/Value[@name="rsv-login-name" and @data="name3"]]/Value[@name="rsv-group-name"]/@data' -nl file.xml 

Or, with a query value taken from the command line:

xmlstarlet sel -t --var data="'name3'" \ -v '//Key[Key/Value[@name="rsv-login-name" and @data=$data]]/Value[@name="rsv-group-name"]/@data' -nl file.xml 

Alternatively,

xmlstarlet sel -t --var data="'name3'" \ -m '//Key/Key/Value[@name="rsv-login-name" and @data=$data]' \ -v '../../Value[@name="rsv-group-name"]/@data' -nl file.xml 

The output from either of these will be group 2.

The XPath query looks for a Key/Key/Value node that has name="rsv-login-name" and data="name3", and for the Key that this node is a Key/Value node of, it returns the data attribute for the Value node that has name="rsv-group-name".

The XML I've used here is a slightly modified document that adds a <root> tag at the start and a </root> end tag at the end, just to make it a well-formed XML document.

Answer 2

Look into usling xslt scripts to transform or parse your XML, xsltproc is the binary you are looking for and w3cschools have a good getting started guide.

Answer 3

awk solution.

$ awk -F= '/name="rsv-group-name"/{g=$NF}/data="name4"/{print substr(g,2,length(g)-5)}' inputfile