query based on url

[jdulberg]

I was wondering if its possible to make a query based on the url that someone accesses a shared database from.

I'm working on a realestate listings database that is shared across multiple domains on the same server.

Here's and example of what I'm currently using.
agentwebsite.com/listings/index.php?agent=2

and this is what I'd "like" to have -- just makes it look nicer.
agentwebsite.com/listings

I don't mind having ?agent=x in all the other URL's in the listings system but for advertising purposes, its nicer to see the main link without a query attached to it.

Is something like this possible? Any suggestions are appreciated.

Thanks.

[qslack]

On a properly-configured server, if you go to /listings/, it'll be the same as going to /listings/index.php. But you do need to include the agent ID somehow.

I may have misunderstood your question, but do you want to get the agent ID from the domain name? For example, do you want domain1.com/listings/ to be agent ID 1, and so on?

[jdulberg]

Yes, the server is set up so it will automatically look for the index file.

Your right on your second assumption. I would like to get the ID based on the domain name.

agent1.com/listings = agent1.com/listings/index.php?agent=1
agent2.com/listings = agent2.com/listings/index.php?agent=2

etc.

Thanks for your time.

[qslack]

Try this:

switch($HTTP_HOST) {
       case 'agent1.com':
              $agent = 1; break;
       case 'agent2.com':
              $agent = 2; break;
      default: # if nothing else matches
              $agent = 3;
} 

So going to agent1.com/listings/ would set $agent to 1, agent2.com/listings/ would set it to 2, and so on.

[jdulberg]

Do the domains need to be hard coded? There's a potential for about a hundred or so agents to be on this thing so adding each site to the .php file would take a bit of time.

Each agent already has a bit of information in a table, if needed, I could add a field for their URL too.

Would something like this work? If so, what else would I need to make it work like your example?

(select id from agents where url ="$HTTP_HOST"); 

I suppose its the same sort of idea as the spider friendly URL's.

Thanks.

[qslack]

Yep, that would work. Then just set $agent to 'id' from the SQL statement.

[jdulberg]

Great, I got it working --- mostly. It only works if $HTTP_HOST matches exactly what I have in the database. So if I have the www. in the db, and don't put that in the URL, it won't work. Is there any way that I can fix this?

$headinfo = mysql_query("SELECT id FROM agents where agenturl='$HTTP_HOST'",$link);
    if (!$headinfo) {
        echo("<P>Error retrieving agent from database!<BR>". "Error: " . mysql_error());
        exit();
    }

    $headinfo_redirect = (mysql_fetch_array($headinfo));
    $id=$headinfo_redirect[id];

if ($id) {
$agent=$id;
} 

Thank you again.

[Marty H.]

You could try

$headinfo = mysql_query("SELECT id FROM agents where agenturl like %$HTTP_HOST",$link);

This would allow for differences at the biggining of http_host. But if http_ host happens to be 'abc.com' it would also return results for cdeabc.com.

Or you could store all hosts as www.abc.com and the test to see if $http_host has www as the first 3 characters and if not then do

$http_host = "www.".$http_host

before doing the query.

[jdulberg]

How would you go about evaluating $http_host to determine whether or not it has a www in it and if not, plugging it in.

Thanks again for your help!

[DR_LaRRY_PEpPeR]

you want to add the "www." if it's not there right? this reg exp should do the trick:

if (!preg_match('/^www\\./i', $http_host))
{
    $http_host = "www.$http_host";
} 

if you want to strip the "www." part if it exists (and just search for the domain), i can show you the reg exp for that if you want that.

[jdulberg]

Yes, that would be great if you could show me how to strip out the www - I think it would be easier to store them without it.

Thanks.

[DR_LaRRY_PEpPeR]

k.

$http_host = preg_replace('/^(?:www\\.)?(.*)$/i', '$1', $http_host); 

if your PHP isn't at least version 4.0.4 you can't use that '$1' part. you'd need it to be '\1'. but '$1' is the preferred way.

[jdulberg]

Sweet! works perfectly. Thanks. phpinfo says that the server is running php v. 4.0.1pl2 so it looks like I'll have to upgrade soon.

Thanks again.

[DR_LaRRY_PEpPeR]

good to hear. no problem.

[jdulberg]

Everything is working fine with the regex that you gave me a while ago but it seems that I need your help to take it just a bit further. I now need to remove http:// from addresses as well. Seems that some people on my site don't follow instructions so I have to make sure the url's are just the domain name using a more detailed regex.

So if possible, I need urls like:

http://www.domain.com
http://domain.com

to be

domain.com

The code that I am currently using is this:
$newsletter = preg_replace('/^(?:www\.)?(.*)$/i', '\1', $newsletter);

I tried to do it on my own but just end up getting confused.

Thanks again for your time

[delemtri]

Augh, regular expressions! I'd just use substr.

[DR_LaRRY_PEpPeR]

oh but del reg expressions are FUUUUUN man!

jdulberg, this should work. i haven't tested it, but give it a try.

$newsletter = preg_replace('#^(?:https?://)?(?:www\\.)?(.*)$#i', '$1', $newsletter); 

that will handle "https" also. if you don't want that, take out the "s?". and don't let the pound signs (#) alarm you - i just use those as the pattern delimiter so i don't have to escape the "/" in http://.

and again, since i guess your PHP version is < 4.0.4 replace the $1 with \1.

let me know if it works for ya.

[jdulberg]

Awesome, thanks for your code!! Its working perfectly. I took out the "s" from http as I'm pretty sure that the people using the site won't be linking to secure pages. Had to put the \1 as you suggested too as I only have php v. 4.0.1

Thanks again for your time!

[DR_LaRRY_PEpPeR]

no problem.

if you took out the "s", did you take out the "?" too? because if you have it as "http?://" that will make the "p" optional. so in other words, you want this in the first set of parentheses: (?:http://)

just makin' sure it's right for ya.

[jdulberg]

ah... good thing you told me Thanks for letting me know! So the final code is now:

$newsletter = preg_replace('#^(?:http://)?(?:www\.)?(.*)$#i', '\1', $newsletter);

Thanks again for your help!

[DR_LaRRY_PEpPeR]

yes, that code is correct. and, again, no problem.