Oh my bad! Instead of using a simple GROUP BY + MIN (), I've been following my strategy for years just because I was unable to use COUNT over grouped results. However, doing some tests now, I've found a solution.

According to what I explained in my previous message, the lack of the DISTINCT ON clause can be avoided using this:

SELECT tbl.F1, MIN (tbl.F2)
FROM tbl
GROUP BY tbl.F1

But, to count the results, I've came to this solution:

SELECT COUNT
FROM tbl
WHERE tbl.ID = ANY (

SELECT tbl.ID
FROM tbl
GROUP BY tbl.F1)

It's MUCH more efficient than my previous strategy.

I feel quite silly, because I've used something similar to solve other lacks of Hibernate. But this idea didn't came to me in this case :/ . Anyway, I'm happy to have a good workaround now .