Grep a word from XML file with grep/sed

Question

I am working on a linux shell script to find information in a xml file using grep. The information I need is the Username wipis_dxu without the tag. I am not allowed to use tools like xgrep or something else.

<Parameter displayName="Server" id="Server">
 <value>dxstg.target_domain</value>
</Parameter>
<Parameter isRequired="true" displayName="User name" id="Username">
  <value>wipis_dxu</value>
</Parameter>
<Parameter isRequired="true" displayName="Password" id="Password">
  <value>wovon_man_nicht_reden_kann_darueber_muss_mann_schweigen</value>
</Parameter>

Answer 1

grep -r "wipis_dxu"

Will search current directory and all sub-directories for "wipis_dxu".

Answer 2

$ sed -nr '/Username/{n;s#.*>(.*)<.*#\1#p}' file
wipis_dxu

Brief explanation,

• /Username/: find Username
• n: Read the next line of input
• s#.*>(.*)<.*#\1#p: extract the string between > ... <

Answer 3

GREP only

This is a bit tricky because grep matches line by line, and on the line of the username there is nothing specifically saying it's the username and not the password.

So you'll have to chain greps :

grep -A 1 "Username" file | grep -o -P '[^>]+(?=</v)'

Answer 4

Don't use sed and grep to handle XML, use an XML-aware tool. For example, in a tool I maintain called xsh, you can write

xsh -I file.xml 'echo //Parameter[@id="Username"]/value'