By default, the driver now uses a new bulk command that executes batch a lot faster. The main default is that connector then just now the total results of the batch.

Example for 3 inserts:
before, connector know that 1 + 1 + 1 inserts where executed.
now, the driver just knows the total results is 3 for the example, so connector return array of -2 (= SUCCESS_NO_INFO).

If update results are important, this new fast batch can be disabled by setting the option "useBulkStmts" to false.

Please read the description carefully. I was talking about UPDATE/DELETE statements, NOT INSERT statements!
Of course, when I execute 3 INSERT statements in a batch, unless this fails with some constraint violation, I know that the total count is 3 (I don't even need to check the result array whether it contains -2 in each bucket).

However, let's say I execute these 3 statements as a batch (by calling PreparedStatement.executeBatch())
UPDATE ORDER SET STATUS = 'DONE' WHERE STATUS = 'IN PROGRESS' (e.g. this one updates 10 rows)
UPDATE ORDER SET STATS = 'CANCELLED' WHERE STATUS = OPEN' (e.g. this one updates 0 rows)
UPDATE ORDER SET STATUS = 'REOPEN' WHERE STATUS = 'IN PROGRESS' (e.g. this one updates 5 rows)
With the new driver and without setting any options, how to I get the individual update counts (10, 0, 5), or at least the total count (15) ??? As I wrote in the description, no usable method returns anything I could do that.
Note that with driver 2.1.0 the getLargeUpdateCount() would return 15, with latest driver it will return -1.
Performance is very important but cannot go agaist usability.

How do I set the useBulkStmts option?

I've just discoverd, by complete chance, that CONJ-642 (fixed in 2.3.0) solves this problem.
Shouldn't this issue be then closed?

closing since solved by CONJ-642