Hi Mr. Dirk,
Yes I did try pushing it down to the database level but the TCRdr is what is taking longer. The throughput is too low. Specifications of our new servers are much higher (CPU, memory) than the previous so I needed to determine which is really causing the performance issue (could it be database parameters/configuration etc). In reference to your response to other thread, can you share how I will be able to gauge below:
There's no theoretical limit. In your case, total elapsed time will depend on 3 parameters:
1/. how fast can your database sort 10 million rows?
2/. how fast can DS read out 10 + 20 million rows?
3/. the number of comparison columns
For now we'd opted to performing optimization since the same process is working fine in the current production server. Optimization is a last resort but I am already doing it, in hopes to find some clues. (By the way, thanks for your articles 
)
Thank you in advance. It is much appreciated.