Kris Jurka wrote:
>
> On Mon, 20 Sep 2004, Kris Jurka wrote:
>
>
>>At the same time do you think we should change the default fetchSize to
>>1000 or so? Even though it only works for FORWARD_ONLY ResultSets it
>>could reduce the number of complaints we see here without much penalty.
>>Previously the cursor implementation was more limiting by ruling out
>>server prepared statements.
>>
>
>
> Actually a little testing shows that a non-zero fetchSize implies using a
> named statement instead of the unnamed statement. So there will be some
> potentially negative planning differences, so I'm going to retract this
> suggestion.
Ah, this is because if we used the unnamed statement, the created portal
gets closed when the unnamed statement is reused. I forgot about that.
-O