> Moving the implementation is out of my hands. > > JDO for RDBMS is an ORM solution which doesn't necessarily > answer the question, why not use both (JDBC and ORM)? I think > both solutions have merit and I think that many try to make > one solution work for both OLAP and OLTP. >
Ok, in this case, how about splitting the databases? Keeping OLTP part as is, and creating the OLAP part on daily (hourly or whatever) import basis as a separate DB, where you can create fast indexes, which you can't afford in the OLTP part. Or do you always need to analyse the data instantly? We once made it for a project, where the customer wanted to create new queries for statistical purposes from scratch. Like, today i'd like to know, whether the 50.000 users registered in last month were more female of male, and how many of male users came from north dakota... We created a table, which was imported every night, and contained all the information about a user, which was available in different parts of the system. So the customer was able to analyse everything, without running queries against the live-system. Regards Leon --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
