Re: millions of records
Kevin:
I am going through something similar as you are. But in my case, I need my database design ready to handle 50 million records ( im shooting for the moon here ).
lets pretend im amazon, and i have 50 million rows of customer data.
obviously, i cant store this in one table can i.
So when i think about it, I would have to setup some sort of system for storing this..
After a lot of thought I figured, how about creating customer id #'s?
for example: bob hunt gets a customer id # 10098994488
for example: jerry fox gets a customer id # 10198994488
I could take the first three digits (100 and 101), and store them, but when you think about it, .... how are you going to look up this customer through 50 million records? When a user comes to your site, you dont ask for a freaking customer id #, your site asks for their email and password. So you have an email address, and password, but no customer id #? so how are you supposed to know which table they are in? And even if you search all .... 500 or more tables, ... with 400,000 lines or more in each table, it would take a lot of time even with indexes and primary keys.
So if we did something like this, where would the entire 50 million rows of data holding the email address and customer id # go so you could search it and get the customer id #, and then go to the table they are stored in.... that's the really screwed question here.
if you figure it out, let me know.