thanks, Mike. for my case, rows in table is more than 5,000,000 . There is a transaction need to insert one row in a few tables like insert into... value.... the sequence ID is generated by customized code.
User request no delay for this transaction but it has kind of delay sometime.
So If I change all primary key as non-cluster, should have improvement for this transaction, right?
As Primary key is unique, for query, it always like where id = id value,=always for one row. So for this case, cluster or non-cluster should have no big difference for this query, right?