Hi Chris,

Thanks for the leads, but even with recompile and changed index, the time is the same.

In relation to your first post, then I will not likely get anymore than a few 1000 rows as result (maybe 10t-20t), as it will not include more than 5 trade lanes and there are around 250.000 different tradelanes in the table. Does that make it easier to optimize for that specific purpose?

Furthermore, where can I see if the CPU or the disk (or something third) is the bottleneck?