Andrew Roth wrote:
Hi all,
We are a group of three students in Professor Ric. Holt's Software
Architecture class at the University of Waterloo. As our project, we
would like to examine the MySQL source to determine the best way to
implement materlialized views.
It would be wise to hang around and see if someone from MySQL AB has
time to answer the questions, but here's my personal view...
I had to look up materialized views and if I read it correctly what is
said about this at
http://www.akadia.com/services/ora_materialized_views.html:
1. How feasible would implementing materialized views be?
It would be quite an adventure, but they might be useful in certain
situations (low concurrency, but need for more speed or higher
concurrency and less need for accuracy).
2. Any reasons why materialized views haven't been added already?
Since 'views' are only introduced in MySQL 5.0 I guess that:
- there has not been time yet to think about materialized views
- no programmer has had the time to build it
- no-one has sponsored to hire extra programmers to build it
4. Any comments at all relating to this endeavor.
Good luck? ;-)
I think that you have to make changes in a lot of areas. At least you
will have to think of:
- storage engine(s) (updating materialized views on commit)
- query optimizer (trying to rewrite the query to use MVs instead)
- adding handling of the new keywords
- sceduling of refreshes during off-peak time
Kind regards, Jigal.
--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
