SET table with UPI is always better because it enforces the uniquencess of the data in turn it ensures that data is distributed across AMPS evenly. A evenly distributed data improves the performance of the retrieval.
As Rajeev rightly mentioned, if you define a NUPI, then the SET operation needs to scan the entire row ensure that there are no row duplicates. Hope this helps.
SET table with UPI is always better because it enforces the uniquencess of the data in turn it ensures that data is distributed across AMPS evenly. A evenly distributed data improves the performance of the retrieval.
As Rajeev rightly mentioned, if you define a NUPI, then the SET operation needs to scan the entire row ensure that there are no row duplicates. Hope this helps.