Currently we have teradata 14 system. I often see a high impact CPU for corelated subqueries. Ex. A query with 4k CPU has ImpactCPU of 10k. My question is how much a coreated subquery efficiently works in teradata. Thanx.
Impact CPU is related to skew. Your correlated subqueries are causing several AMPs work more than others.
What kind of Correlated Subqueries?
Scalar within the SELECT list?
Using non-equality conditions?