DSXchange

i am doing a join on input data from vitual dataset and an oracle table. The key is a single column of integer type. Right now i am hash partitioning and sorting both the inputs. Instead of Hash can i use Modulus. Developer guide says YES. But if i use modulus, does it ensure related records stay in same dataset. and which is fastest, hash or modulus. How the modulus work. does it also create any table like hash

Modulus is more efficient than Hash because it does less work. Do an all terms search for both terms; I've explained it elsewhere. Modulus algorithm does guarantee key adjacency on each partition.