Bash scripting - string searching

[rampant]

I am trying to extract text from a file, say the file looks like this:

</tr><tr><td>domain1.com</td><td>2006-11-06 14:43:37</td>

<td><a href="index.php?_host_id=47">1</a></td>
<td><a href="index.php?_host_id=47">2</a></td>
<td><a href="index.php?_host_id=47">3</a></td>

</tr><tr><td>domain2.com</td><td>2006-11-06 14:43:37</td>

<td><a href="index.php?_host_id=48">1</a></td>
<td><a href="index.php?_host_id=48">2</a></td>
<td><a href="index.php?_host_id=48">3</a></td>

and many other lines like this. What I need is a way to get all of the host_id's only once and only 1 per line, like this:

47
48

I have found the domain string with this already:

awk '/domain/' file | cut -c 15-25

[Javasnob]

I would look into grep and sed.

[Marsolin]

If all the lines follow that same format then I one possibility would be the following:

grep '_host_id=' file | cut -d '=' -f2 | sed 's/\".*$//'

grep only returns the lines with the "_host_id" string
cut returns the text following the second = sign
sed strips everything from the " and beyond

You could probably get rid of cut and do the line processing with just sed, but I don't have the time to figure it out. I also haven't tested this so it may be slightly off.

[rampant]

Figured it out, thanks for the help.

Didn't think to just 'cut' it up.

Something useful too:

'sort -um' will give you unique instances of strings within a file without sorting the results by value.