Possible encoding problem in sql queries

Question

I am using mac os. I've set everything possible (files, eclipse workspace, database collation, mysql charst) to utf-8 encoding. Still when i send a query, smthing like
SELECT * FROM suppliers WHERE LOWER(name_post)='акцент'
using java, empty resultset is returned (it should be nonempty). My only guess is that cyrillic part of the query is not correctly encoded. I used to have lots of those queries under windows w/o any problem. Is there any way to figure out what is going wrong? (If I select the whole table and then verify the constraint within a java loop it works fine)

Answer 1

If you use a Statement, and concatenate your parameters in like:
String query = "select * from my_table where some_id = " + anId;Oracle will create an execution plan for each id that passes through this bit of code. This execution plan gets stored in the shared pool. The more of these that happens, the fuller the pool gets. Oracle will age out older plans, but this can result in pool fragmentation if there's a lot of "churn" in the pool, which there will be if your queries look like the above.
Eventually you'll hit an ORA-4031, as it can't find a suitable chunk of pool to allocate to a query. Some people see that as "I've run out of pool" and so try to "fix" it by expanding the shared pool size. All this tends to do is delay the inevitable.
What a 4031 actually means is "you're not using bind variables", since binding would result in a single plan being held for the above SQL, since it would actually be stored as "select * from my_table where some_id = ?". Thus no flooding of the shared pool.
So, PreparedStatements aren't just about security, but also about good resource management.
Hopefully that makes sense?