[email protected] wrote:
If you try to keep the historical temporal record inside the SQL tables for the
running application it is going to cause an EXPLOSION in the number of records
in the tables. Why? Because every change that you make, instead of being an
UPDATE, becomes and INSERT with a new timestamp. It also makes the raw table
records much harder to retrieve and read using SQL in terminal windows.
We explored and experimented with all of these ideas years ago. The idea
behind using a second set of tables was so that you did not impact the volume
or performance of the running application tables. That's why every company
where we were involved with these efforts used a second set of tables.
Yes, you will have many more records, but aren't todays RDBMS capable of
handling millions of records with ease?
Premature optimization maybe?
_______________________________________________
Tinyerp-users mailing list
http://tiny.be/mailman/listinfo/tinyerp-users
