>-----Original Message-----
>From: phobo [mailto:[EMAIL PROTECTED]]
>Sent: Tuesday, January 23, 2001 9:48 PM
>To: [EMAIL PROTECTED]
>Subject: Re: [PHP-WIN] Random Code
>
>
>Oops, 36^8 :)
>
>This value would only have to generated perhaps a few times an hour, but
>should't take more than a second to compute. These values ARE being stored
>in a database; however, once this database gets to a million or so rows, it
>would take a very long time to check it doesn't overlap, wont it?
I don't belive it will take seconds to look up the result (that would be
very bad performance of the RDBMS then...) If you just index the data
properly, then most time probably will be spent transferring the answer over
the network (that is if you use a network:). If your time constraints is
within a second, and you do not need to generate the numbers more often than
a few times in a hour, then why not do it?
I have a table (among others) with 3.368.713 rows in my database. This is an
execute example of an look up onto an attribute with an index in the
database:
I executed this query:
select getdate()
select count(*)
from luAccNr
where AccNr = 'X02612'
select getdate()
The result:
--------------------------
2001-01-26 12:14:24.717
(1 row(s) affected)
pkAccNr AccNr Speice
----------- ---------------- ------
629888 X02612 1
(1 row(s) affected)
---------------------------
2001-01-26 12:14:24.763
(1 row(s) affected)
As you can see, the execution time for the query is not even a question of
tenth of second, but down to about 50ms. And this includes to execute the
getdate() functions too.
Best regards,
Anders
--
PHP Windows Mailing List (http://www.php.net/)
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
To contact the list administrators, e-mail: [EMAIL PROTECTED]
