controlling number of SQL operations

[Adam Hardy]

I am running a load test for a performance-sensitive operation and I'm surprised 
by the number of SQL calls I am seeing, so I want to check that I have the JPA 
and OpenJPA config set up right.

Hopefully someone can say if any of the SQL calls are avoidable.

This is the entity model:

CollatedAnalysis -> Trade -> Balance

The operation, CollatedAnalysis.analyze(), causes the entity to run over its 
large collection of Trade entities, producing a new 'grandchild' Balance entity 
for each Trade.

What happens in terms of SQL hits is this:

select all Balance where Trade = ? (because I put the Balance in Trade.balances)

insert Balance

update Trade



For the load test I run the operation on 30K trades - naively I thought I would 
only see 30K inserts - is there any way to avoid the other 2 operations?


Thanks
Adam