[DB-SIG] which free database to use (and a few other questions)

aaron watters aaron_watters@classic.msn.com
Wed, 21 Jan 98 20:52:33 UT

I betcha you folks knew this was coming, dincha?

You could also use gadfly: pure python sql.  I don't know about how the 
performance would compare.
I'm pleased with its performance myself, especially if you install kjbuckets.

One problem: it's 100% in memory at present so you'd need lots of RAM and huge 
swap files (500M)
But that's perfectly doable (also you'd best use kjbuckets, since it uses 
memory better).

Realistically, even if you had Oracle you'd probably want to 
preprocess-extract the data first, if possible.
100MB is pretty huge. Anything is likely to be very slow.

Alternatively you could subclass the "Relation0" class to read from disk in 
some cases.
Easily done, actually, but I don't know how to hook such a thing into SQL 
syntax nicely...
If you want to play with it I'd be happy to help (at least a bit).

Alternatively, you could punt on sql and "home grow" your indices and queries 
in Python.
Maybe using bplustree.py.


Just a thought.

New release always rsn.  ;c) -- Aaron Watters

-----Original Message-----
From:	owner-db-sig@python.org  On Behalf Of Cary Collett
Sent:	Tuesday, January 20, 1998 1:24 PM
To:	jeremy@CNRI.Reston.Va.US
Cc:	db-sig@python.org
Subject:	Re: [DB-SIG] which free database to use (and a few other questions)

> I'm looking for some advice on which database (and which database
> module) to use.  I need a free database for several data
> collection/analysis projects, including TCP packet trace and HTTP
> server log analysis.  A typical data set will be 50-100 MB.
> It looks like I can choose between miniSQL and MySQL, but I don't how
> good either product is.  Can anyone recommend one over the other?
> I've spent an hour or two looking at miniSQL and it seems to provide
> adequate functionality (but I don't feel like I have a particularly
> informed decision).

>From a feature and performance standpoint, MySQL beats mSQL hands
down. I started using mSQL and later moved to MySQL and have been
much happier.

I think the user community for MySQL is more active as well, but I
haven't really kept up with mSQL, so who knows.

> I'm also curious about the state of the Python modules for each of
> these databases.  The miniSQL module doesn't conform to the DB API.
> Is this a problem?  It looks like it would be easy enough to use the
> mSQL C library to implement an extension module.  (But the fact that
> it looks pretty easy makes we wonder why no one else has?  Perhaps
> there are some problems with mSQL...)
> I started looking at the MySQL module and stopped right away; the
> copyright notice is downright scary: "Note that any product, process
> or technology described in this software may be the subject of other
> Intellectual Property rights reserved by FGA bitart and are not
> licensed hereunder."  Doesn't sound like free software to me.
> On the other hand, maybe MySQL also has a good C library that would
> make it easy to do a new extension module.  Any thoughts?

You could also use ODBC to talk to MySQL if you're using Pythonwin.

MySQL's C library is very similar to mSQL's and the MySQL module
is based off of the mSQL module for this reason. IANACP (I am not a
C programmer), but the API looks perfectly reasonable to me.


DB-SIG  - SIG on Tabular Databases in Python

send messages to: db-sig@python.org
administrivia to: db-sig-request@python.org

DB-SIG  - SIG on Tabular Databases in Python

send messages to: db-sig@python.org
administrivia to: db-sig-request@python.org