By the way, 10s is still a long time to read 10,000 rows, but I suspect that 
the delay in this particular case is primarily due to the fact that the JDBC 
call is being made over a VPN connection. In other words, it's not endemic to 
the ResultList class itself - reading the data from ResultSet directly produces 
similar results.


On Aug 27, 2010, at 6:44 PM, Greg Brown wrote:

> FYI, the change to ResultList does remove some functionality, but 
> significantly improves performance. A query that had taken 75-85s to process 
> ~10,000 rows on a DB2/400 database now takes 10s.
> 
> The major change is that the ResultList iterator now only supports 
> forward-scrolling behavior. However, since that is the common case (and the 
> case for which the ResultList class was originally written), I think this 
> loss is negligible compared to the performance gain. Hope you agree.
> 
> G
> 

Reply via email to