I am just investigating the choices to increase performance of my Data Minning and analisys tool.

I have MYSQL Data base with millions of records. I extarct these records in a Java Swing Graphic Application and store in a table.

Now, when I plot graphs over these records I go to data base again and fetch records set that match what I want to plot.

How can Hibernate fit into this? Can I map my data base tables to objects through Hibernate and run query to those Hibernate objects rather then going to data base again and again with my graph criteria? Will these Hibernate ORM allows data base tables to be in memory and allow SQL to be run on those objects?

Please provide your thoughts.

Thanks
Lucky

Yours doesn't sound like a good application for ORM. You should probably use a cache of some sort, like ehcache, directly.

[Lucky75] That is what I am trying to figure out if I should use Hibernate or not. In the presentation Hibernate states that it supports "high performance queires" as well as "Dual Layer Cacheing Architecture".

What I am thinking is map my tables to Hibernate objects and use HQL on those. But I am not sure if Hibernate is fast enough to execute my queries on millions of records OR NOT. It does have ehcache support built in though.

Any thoughts?

Thanks

Modern databases have their own caching schemes. Hibernate probably will not give you any magic speed improvements over correctly tuned raw database queries.

It's possible, though, that if you have a lot of RAM and set up a large cache, you may see some worthwhile improvements.

The only way to know for sure is to set up a test case for each approach and measure their speed.

There are special kind of features in databases aka "OLAP" ( bitmap indexes an mviews, special kind of statistics functions/procedures ). I am afraid "Data Minning and analisys tool " with Swing stuff will fail to load analisys databe to memory.