Question about db abstraction layer

Anon

Im curious how the various db abstraction layers that are set up like Tim's Database Abstraction Layer handle multiple database connections within 1 page? As far as I can tell there is nothing in the code that helps a db_query() call determine which conection to use, short of just useing whatever was the last globally set information. I like the simplicity of those types of abstraction layers, but this seems a little dangerous (read: I might rather go OO than do it that way)

Anon

Actually, I prefer to abstract the database with a very thin layer, so I can change database engines just by changing a line in the connect.inc file I'm using.

I just write a wrapper for the standard connect, execute query, and get a row parts to abstract out the database, so I can handle multiple connects (don't need to often, but sometimes.)

I.e.

$connect=db_connect($server,$user,$etc);
$result=db_exec($connect,$query);
$row=db_fetchrow($result,$index);

etc.

Anon

there's a killer db abstraction layer, that handles multiple connects (among other features), called ADODB... you can find it at
http://php.weblogs.com/ADODB_Manual

best,
Eric

Anon

I have looked at adodb but it still seems to require the hard codeing of db info on a page level, which is something I would like to avoid. I have whipped up a quick example of a class based system that defines the db connection info centrally. Its <a href="http://www.phpbuilder.com/snippet/detail.php?type=snippet&id=146">posted in the shared code lib</a>.