Regex Matching

[EndianX]

This should be a very simple question but I can't find the answer anywhere.

I want to perform a regex match but only have the match returned, not the line containing the match. For example, consider the following:

cat webpage.html | awk '/<title>[^<]*<\/title>/'

That will return the line containing the title element, but I just want the match, not the whole line. How can I accomplish this using awk, or some other standard utility?

Thank you,

Dave

[hazel]

It is possible in awk to use a regular expression as a field separator. I can't remember the syntax for defining it, but if you used \<.*title\> it would turn the title into an awk field which you could then print out.

[WuNMOjTIQanXWBcH]

Attention: the regular expression in the following answer is simplified for the sake of readability.

I want to perform a regex match but only have the match returned, not the line containing the match. For example, consider the following:

cat webpage.html | awk '/<title>[^<]*<\/title>/'

That will return the line containing the title element, but I just want the match, not the whole line. How can I accomplish this using awk, or some other standard utility?

hazel already gave the right hint by saying:

It is possible in awk to use a regular expression as a field separator. I can't remember the syntax for defining it, but if you used \<.*title\> it would turn the title into an awk field which you could then print out.

The correct syntax would be:

cat webpage.html | awk -F '</?title>' '/(<title>.*<\/title>)/ { print $2 }'

Alternative solutions (which totally depend on what you define as a “standard utility”) are:

cat webpage.html | grep '<title>' | sed -e 's/^.*\(<title>.*<\/title>\).*$/\1/'
cat webpage.html | perl -ne 'print "$1\n" if /(<title>.*<\/title>)/'
cat webpage.html | grep -o '<title>.*</title>'

[fiomba]

Hi Endianx,
I often encounter this type of problems, but i use "alternative" solutions

cat xxx.html|grep -i '<title'|cut -d'<' -f2|cut -d '>' -f2

With such a shell expression, you get "only" the title (w/out tags) and
it is working even if the tag has parameters.
A bit of explanation:

                          
cat xxx.htm                  = list the html to be treated
grep -i '<title'             = grep (w/out considering  the case) the tag (w/out '>')
cut -d'<' -f2|cut -d '>' -f2 = isolating the title (-d '<' or '>') as the 2nd field (-f2)

Bye.

[EndianX]

Thanks all.

These all look very helpful. I like the sed and perl solutions. And somehow I'd never even heard of the cut command before.

Thanks again,

Dave