Re: Denormalized for Performance
- From: Juan Carlos Reyes Pacheco <juancarlosreyesp@xxxxxxxxx>
- To: jkstill@xxxxxxxxx
- Date: Wed, 9 Nov 2005 08:37:12 -0400
Hi Jared
We had a sistem similar to a bank, when the customer bough a "share"
You have a table saving the transaction more than 500,000 records and
a table about 3,000 records with customer; in the customers table
there is columns saving the balance, but anyway we had problems,
because there are several transactions, and some complex; so at the
end of the day there is a query validating the customer table column
showing the balance is correct with the transaction table.
Aditionally we had an historic of the table customers, so if you want
to know the balance for a customer on x day, you go to the historical
table (date+customer code) and you find the balance at that day.
The columns for customers, are the usual custoemr code, some personal
data like telpehone, etc. and the balance column.
The columns for transactions are datetime customer, amount in money,
amount in shares, user who did the transaction, detail, type of
transaction and a correlative.
:)
--
//www.freelists.org/webpage/oracle-l
Other related posts:
