SED multiline pattern matching

AutoC · 07-17-2009, 11:26 PM

Hi,
I have an xml file that has multiple entries like..

Quote:

<recording audio="011/011o031e.sph" name="1e">
<segment lang="en_US">
<orth>
Midlantic ,comma a New Jersey bank stock ,comma was up three
quarters to forty seven and three eighths .period
</orth>
</segment>
</recording>

I need to parse each entry, create a file based on audio="",for eg. I need to create 011/011o031e.t and write the stuff between <orth> and </orth> in the file.
any help?

ghostdog74 · 07-18-2009, 12:04 AM

awk

Code:

awk '/audio/{
 for(i=i;i<=NF;i++){
    if($i ~ /audio/){
        gsub(/.*audio=\"|\//,"",$i)
        gsub(/\..*\".*/,"",$i)
        filename = $i".t"
    }   
 }
}
/<\/orth>/{f=0;}
/<orth>/{f=1;next}
f {   print $0 > filename }' file