[REBOL] Re: REBOL and database stuff
From: mat:plothatching at: 30-Mar-2003 19:13
Hello Gregg,
GI> Have you looked at DocKimbel's MySQL interface. People seem to like
GI> that a lot, though I've done nothing more than read the docs on it
GI> myself.
I actually stumbled across it yes. I just thought since I have a
REBOL/Command license that might be preferable.
Certainly not too late for me to shift!
GI> Here's an example for Doc's MySQL:
GI> author: "Elan"
GI> title: "Rebol Official Guide"
GI> insert db ["insert into mybooks values (?,?)" author title]
Yeah same sort of thing as REBOL/command. Identical actually.
Hmm, so... how's he talking to SQL. Does this also require /command?
MB>> Unfortunately one doesn't appear to be able to substitute 'table1' for
MB>> an ? there, so you can't formate a specific table request.
GI> What about using COMPOSE, or REPLACE with a template query string, to
GI> create your SQL statement? Not sure why it might have choked on the
GI> one you created.
You know on reflection, that may have been my fault. It ought to work
shouldn't it? So I'll retry that me thinks.
MB>> Maybe I should have just slapped all my data into one big table rather
MB>> than having many. A little counter-intuitive to me though.
GI> You mentioend having 5,000 fields in your first message. What the heck
GI> do you have in there!? ;) Can you provide an idea of what fields are
GI> important, how many records it needs to handle, and how you need it to
GI> work (e.g. do you get lots of updates from somewhere, mostly filtering
GI> to send groups of emails, etc.)
Right this is my fault for being a total database weenie. I meant
5000 records, not 5000 fields. There's not much in the way of actual
fields. Just name, e-mail, and various fields which have to do with
tracking the user, whether their e-mail has bounced and whether
they've asked to be unsubscribed.
I guess that's not really important now we've established the rows
don't have 5000 fields :)
GI> What would you want in a native REBOL database?
Nothing that's not in that lovely db.r really. It's just that it'd
probably get very slow or break with 5000 records.
My immediate requirement is 3000 records and I had set about doing
that with db.r and suspected it would work. But I thought better to do
the work now, for when I need to do one later in the month with 50,000
records.
I also thought it'd be pretty fast. However just inserting a new row
into a datbase with less than ten fields... it's only doing 2-3 a
second. Can that be entirely down to the MySQL server being on another
machine? Surely it's faster than that?
Regards,
Mat Bettinson - +44-(0)20-83401514.
