[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Subject: Re: [Fwd: RIM 0.54 and RS 0.83 distribution]
Farrukh wrote: > Correct me if I am wrong but the subset SQL syntax should be > realtively easy to map to an Object store that simply stores objects > as defined by RIM (rather than as defined the document DTDs mapped > from RIM). Unfortunately, I believe you are wrong. I think you would need to eliminate arbitrary joins and limit joins to those that traverse a defined relationship. Stating this restriction precisely may be tricky to do elegently. With arbitrary joins, you would need to have a back end system that does SQL query optimization, which is not something that object databases do (certainly ours doesn't). From listening to others in this thread, it seems that allowing arbitrary joins also runs into trouble with DBA's, who want to have a clearer idea of what queries might be run against their systems. Without arbitrary joins, using SQL at all makes much less sense. Instead it would make more sense to do something that looks like filters, where terms in the filter expression may walk relationships using a dot notation (e.g. managedObject.package.name). Michael
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Powered by eList eXpress LLC