RE: [PEAR-DEV] Re: one abstraction layer (my proposal for)
| From: | Marc Boeren | Date: | Tue, 04 Dec 2001 17:04:58 +0000 |
| Subject: | RE: [PEAR-DEV] Re: one abstraction layer (my proposal for) | ||
| Groups: | php.pear.dev | ||
| Request: | Send a blank email to pear-dev+get-3361@lists.php.net to get a copy of this message | ||
Hi,
> A side note: I emailed the dbx folks
Yep, they're here now (that is, I am :-)
> most of our C core db abstraction in place .. currently they only
> support MySQL and PostGreSQL going by the php manual).
That must be an old version (even though that should state ODBC as well),
MS-SQL, Frontbase and ODBC are also supported.
I've exchanged a few mails with Lukas, and I'll quote them beneath for the
interested reader.
Cheerio, Marc.
--8<--------
> Lukas
me
------------
> great .. I must admit I never tested dbx since dbx never got
> far enough in terms of portability (feature and rdbms wise).
I know that. It is just a very simple wrapper, that makes any database easy
to use (no looping through records and such), as well as providing database
abstraction on a minimal level... that is, a common interface to the
abstracted databases. For db-portability you must still create your sql
db-independent, dbx does not support you in this (but since I'm used to it,
I keep my queries simple and portable).
So, if you're looking for a full-featured, true db-abstraction, dbx is not
for you. If you are however looking for a common interface to databases and
queries, you might consider it.
One of the features could be an sql-rewrite-rule (for handling quotes, limit
and such), another could be datatype-abstraction... but most features are
operations on the resultset, such as additional filtering, joining of
resultsets, extracting columns and such. Transactions are on the list as
well, as is an unbuffered-query version for large resultsets. It's just that
there never seems to be time to work at that list ;-)
> But if dbx could atleast do the core chores of a db abstraction layer
> (which would require that it supported a much greater number
> of db's) we could have a very fast core with all those extra portability
> features in php.
Well, it supports 5 databases so far, and if I have a working setup of any
other database that is supported by php, I can add that to dbx in less than
2 hours...
> That would give this db abstraction layer a major performance boost!
Yes, but my company is waiting for paying customers that want support for
another database, before I am allowed to code that... during working hours,
of course. But I don't have an Oracle setup at home... :-)
> So take a look in the archives over on pear.php.net
> The discussion sort of flared up there again ... it did not
> really move as far as I can tell atleast.
I did look at the last 6 or 7 days... no much that is interesting... I did
see some good bits on Metabase and PEAR::DB, but I don't really know
either...