RE: [PEAR] MDB RC1
| From: | Lukas Smith | Date: | Tue, 13 Aug 2002 20:23:30 +0000 |
| Subject: | RE: [PEAR] MDB RC1 | ||
| References: | 1 | Groups: | php.pear.dev php.pear.general |
| Request: | Send a blank email to pear-dev+get-8403@lists.php.net to get a copy of this message | ||
Hi,
> From: Wolfram Kriesing [mailto:lists@kriesing.de]
> i had a quick look into it and i think it will be a great extension
(or
> replacement) to what pear::db currently is able to do.
> just one question: is there not some kind of xml-data spec from the
w3c,
> which also fits for storing db-structures as opposed to the (i think)
> proprietary format used in mdb?
Well the reason that the format ist he way it is simple: it is the
Metabase format :-)
In the future I want to make this all more modular.
Especially the dumping is ugly since the tags are hardcoded into the
manager.php script.
Once this is separated just like in parser.php things will get much
easier to work with different standards. Metabases format has been
around for quite some time now and seems to be working well for a lot of
people.
I am not really aware of any efforts that have a large enough scope.
Most projects focus on just being able to pass data from and to an RDBMS
server in XML format.
But none of these formats deal with creating schemas in xml (with
information like types, length etc).
This is the output that mysqldump gives back when using the --xml
switch:
<?xml version="1.0"?>
<metapear_test_db>
<numbers>
</numbers>
<real_funky_id_seq>
<row>
<sequence>66</sequence>
</row>
</real_funky_id_seq>
<test>
<row>
<test_id>1</test_id>
<test_name>test0r</test_name>
<test_date>2002-02-12 16:33:53</test_date>
</row>
<row>
<test_id>2</test_id>
<test_name>gummihuhn</test_name>
<test_date>2001-02-12 16:34:03</test_date>
</row>
</test>
</metapear_test_db>
Not really helpful, especially if you want to emulate sequences etc.
But there are some thing here where the current MDB/Metabase schema
format could be made to me more similar.
Mainly by dropping the <name> tag in favor of tag that contains the
name.
Also the <initialization> section could be redone to match the above
format.
I think this needs to be well thought out and I assume that Manuel had
good reasons to do things the way he did back then.
> great work :-)
>
thx :-)
regards,
Lukas