Pedro Alves <pmalves(at)think(dot)pt> writes:> It's the same even with statistics=1000.
Uh, you did actually ANALYZE the table after each change, right?Doesn't the EXPLAIN output change at all?
regards, tom lane