use sed to extract a matching number

[esolve]

I have a xml file with a 1 line long string
I want to extract the number after "<yt:duration seconds='" and before "'/>"
like the following example, I want to get 131

.....<yt:duration seconds='131'/>....

the "...." here means the substrings before and after "<yt:duration seconds='131'/>", which are long with different characters, numbers and marks.

how to use sed for extracting the matching number?
thanks

[Irithori]

Imho, a regex is not the right tool for xml files. It is unnecessarily complex and errorprone.

But xml is structured by design.
My suggestion would be to use a tool, that can handle xml natively, such as the ruby gem nokogiri

[esolve]

I know, but for my xml file, it is ok to use sed, I think

[histrungalot]

This works on my Mac:

$ cat YourTestFile.txt
<td><....><href="URLs goes here">blah</...> <td><...><href="URLs goes here">blahblah</...> <yt:duration seconds='131'/> things fdaakda daf
$ sed -e "s/\(.*<yt:duration seconds='\)\([0-9]\{1,\}\)\('.*\)/\2/" YourTestfile.txt
131
$