Hi,
I've recently tested Derby with a SELECT statement containing 5000
elements inside an IN clause. The query looked something like this:
SELECT field1 FROM table1 WHERE field1 IN (0, 1, 2, ..., 4999)
This failed with a server side StackOverflowError. True, I got past this
issue by starting Derby with the "-Xss1024k" JVM option, but this was a
trial and error approach.
Are there any recommendations on how to set the size of the heap and
stack for a Derby instance? Is there a way to estimate the memory
requirements of a Derby database? Did anyone had such experiences?
Regards,
Robert
PS: I've been using the Derby 10.2.1.6 release in networked mode running
on Sun JRE 1.5.0_09.
