Hard time writing SED or Regular Expression to isolate string

[nauliv]

Hello Folks,

I'm trying to parse an HTML file to extract a city name, using bash scripting on Linux. The tags around that city name are always formatted the same way. Here is an example with LOS ANGELES:


<....bunch of HTML code.....>
<tr><td align="right"> Prefix</td><td><b>240</b></td></tr>
<tr><td align=right>City</td><td><b>LOS ANGELES </b></td></tr>
<tr><td align=right>State</td><td><b>California</b></td></tr>
<....bunch of HTML code.....>


Basically, I am looking into isolating the string between

[highlight]City</td><td><b>[/highlight]

and

[highlight]</b></td></tr>
<tr><td align=right>State[/highlight]

I have banged by head trying to do a 2 process job:
* remove everything before the city name
* remove everything after the city name
So I end up with what I need... but I am sooooo getting lost in the sed and RE expressions.... :-(


Any help or hints would be greatly appreciated !

 

[Annihilannic]

Try this:

sed -n '/City/{ s/.*<b>//; s/ *<\/b>.*//; p; }' inputfile

Annihilannic.

 

[nauliv]

Hello Annihilannic,


Thanks so much for your fast reponse. ALMOST THERE! There's just some extra code after the city name.

Here is the result:

[highlight]LOS ANGELES
<A href="ZipCityPhone.asp?90071">90071</A>
<A href="ZipCityPhone.asp?90013">90013
<A href="ZipCityPhone.asp?90014">90014[/highlight]

You can see the source of the following URL:

http://www.melissadata.com/lookups/phonelocation.asp?number=2139743211

My script is:

#!/bin/bash
callerid=2139743211
curl -s -m 3 -A Mozilla/4.0

http://www.melissadata.com/lookups/phonelocation.asp?number=${callerid}

| sed -n '/City/{ s/.*<b>//; s/ *<\/b>.*//; p; }'

Thanks again !

 

[Annihilannic]

Just widen the "City" match so it doesn't match ZipCityPhone... i.e. include the greater than/less than signs either side of it.

Annihilannic.