Parsing html content with grep

[Stonecold1995]

I have a html page things like this in it:

<a id="p1874298" href="index.php?page=post&amp;s=view&amp;id=1874298">

How do I use grep to output only the number (i.e. "1874298")? So far this is what I have:

wget -O - example.com | grep -m 1 --only-matching --perl-regexp '<a id=\"[^<>]*'

But that doesn't fully work...

[markbl]

wget -O - example.com | sed s'/^.*id=\([0-9]*\).*$/\1/'

[Stonecold1995]

So then, would this be the best way to do it?

wget -O - example.com | grep -m 1 --only-matching --perl-regexp '<a id=\"[^<>]*' | sed s'/^.*id=\([0-9]*\).*$/\1/'

Isn't there a better way do do this, rather than using grep and sed? I'm sure it can be done with grep alone.

[markbl]

So then, would this be the best way to do it?

wget -O - example.com | grep -m 1 --only-matching --perl-regexp '<a id=\"[^<>]*' | sed s'/^.*id=\([0-9]*\).*$/\1/'

Isn't there a better way do do this, rather than using grep and sed? I'm sure it can be done with grep alone.

Why did you add that grep back into the command I posted? You only need sed.

[markbl]

I just realised you probably want to filter other lines out and just grab the number[s] from all the input. So you want something like:

wget -O - example.com | sed -n '/<a id=/s/^.*id=\([0-9]*\).*$/\1/p'

[Stonecold1995]

So then if I wanted to only output the FIRST number, would it be best to pipe that to "head -1"? Or does sed have a feature like grep's "-m 1"?

[markbl]

Yes, If I wanted only the first line of output then I would just add a "| head -1" on that sed line.

However, if you want a more robust solution to parse xml/html tagged data from the command line then I would look at using xmlstarlet. I have used xmlstarlet previously and it is excellent.