Package io.trino.sql.planner.iterative.rule

Class TransformCorrelatedGlobalAggregationWithoutProjection

java.lang.Object
io.trino.sql.planner.iterative.rule.TransformCorrelatedGlobalAggregationWithoutProjection
All Implemented Interfaces:
Rule<CorrelatedJoinNode>
public class TransformCorrelatedGlobalAggregationWithoutProjection extends Object implements Rule<CorrelatedJoinNode>
This rule decorrelates a correlated subquery with: - single global aggregation, or - global aggregation over distinct operator (grouped aggregation with no aggregation assignments), in case when the distinct operator cannot be de-correlated by PlanNodeDecorrelator It is similar to TransformCorrelatedGlobalAggregationWithProjection rule, but does not support projection over aggregation in the subquery

In the case of single aggregation, it transforms: 


 - CorrelatedJoin LEFT or INNER (correlation: [c], filter: true, output: a, count, agg)
      - Input (a, c)
      - Aggregation global
        count <- count(*)
        agg <- agg(b)
           - Source (b) with correlated filter (b > c)
Into: 

 - Project (a <- a, count <- count, agg <- agg)
      - Aggregation (group by [a, c, unique])
        count <- count(*) mask(non_null)
        agg <- agg(b) mask(non_null)
           - LEFT join (filter: b > c)
                - UniqueId (unique)
                     - Input (a, c)
                - Project (non_null <- TRUE)
                     - Source (b) decorrelated

In the case of global aggregation over distinct operator, it transforms: 


 - CorrelatedJoin LEFT or INNER (correlation: [c], filter: true, output: a, count, agg)
      - Input (a, c)
      - Aggregation global
        count <- count(*)
        agg <- agg(b)
           - Aggregation "distinct operator" group by [b]
                - Source (b) with correlated filter (b > c)
Into: 

 - Project (a <- a, count <- count, agg <- agg)
      - Aggregation (group by [a, c, unique])
        count <- count(*) mask(non_null)
        agg <- agg(b) mask(non_null)
           - Aggregation "distinct operator" group by [a, c, unique, non_null, b]
                - LEFT join (filter: b > c)
                     - UniqueId (unique)
                          - Input (a, c)
                     - Project (non_null <- TRUE)
                          - Source (b) decorrelated