I have a file (xml) with the following content:

$cat XXX.xml

<?xml version="1.0" encoding="UTF-8"?>
.
.
.
<subject>
<subjectIdentifier>127_S_5218</subjectIdentifier>
<researchGroup>Patient</researchGroup>
<subjectSex>F</subjectSex>
<subjectInfo item="DX Group">SMC</subjectInfo>
.
.
.



*I want to grep the value specifying the Dx group. In this case SMC. This value will change between different xml files...

So I tried:
$cat XXX.xml | grep "DX Group"
<subjectInfo item="DX Group">SMC</subjectInfo>

But it grep the entire line rather that just the characters i need (e.g. XMC).
Any help?

firstly avoid useless use of cat
http://partmaps.org/era/unix/award.html#uucaletter

now to achieve what you want
a few ways,
awk is useful here

there we use both < and > as field separators

and print "field 3" of lines that contain "DX Group"

grep is for finding text - even has the option to only print what was found (rather than the entire line). Doesn't really fit for this - sed is another option

Great..thank to all, I'll give all the suggestions a try!