When I worked on a project using Hibernate, it generated lots of realy ugly queries, often joining 30 or 40 tables. If a web page required data from several tables, it would try to get all the data in a single query, even if there were many intermediate tables. If it could find a relationship, no matter how remote, it was going to use it.
I'm told that this is done when one chooses the "eager" mode; there's another "lazy" mode that would load the child tables data only at access time. Of course we'll investigate further - but I could make good use of any early advice or pointer to any good doc about Hibernate features for the Oracle perf tuner / re-designer. About the suggestions about using 10046/statspack - I was planning to do that (and the Java guys are already compiling SQL logs using their own tools) - but it's important IMHO, before jumping headlong in tkprof/10046 etc, to get an high-level understanding of the architecture of Hibernate, again in relation to Oracle. Anything can (and will) help. -- Alberto Dell'Era "Per aspera ad astra" -- //www.freelists.org/webpage/oracle-l