Ali Alsuliman has posted comments on this change. Change subject: [ASTERIXDB-2015][IDX] Introduce Primary Index Optimization Rule ......................................................................
Patch Set 8: (6 comments) https://asterix-gerrit.ics.uci.edu/#/c/2111/6/asterixdb/asterix-algebra/src/main/java/org/apache/asterix/optimizer/rules/am/IntroducePrimaryIndexForAggregationRule.java File asterixdb/asterix-algebra/src/main/java/org/apache/asterix/optimizer/rules/am/IntroducePrimaryIndexForAggregationRule.java: Line 60: * ^ > It seems to me that an aggregate followed by assign is really not a prerequ Yes, aggregate->assign is not a prerequisite. That's why its existence is optional. Assign operator will be there for the use case: count(*) where the query plan has assign $ = 1. Originally, the optimization was for count(*), but I wrote the rule to work for aggregations in general (involving only PKs) Line 91: */ > Let's make field type List<> and use ArrayList<> as implementation, so we a Done Line 139: private Pair<Mutable<ILogicalOperator>,Mutable<ILogicalOperator>> findScanAndAssignOperator( > I think an aggregate operator can only have one input. Done Line 156: UnnestMapOperator unnestMapOperator = (UnnestMapOperator) scanOpRef.getValue(); > Use AbstractFunctionCallExpression here instead of UnnestingFunctionCallExp Done Line 195: dataset.getPrimaryKeys().size(), scanOperator.getVariables().size()); > If you're replacing index-search() then I think you can just get requiresBr Done Line 208: scanOperator.getInputs().get(0).getValue().getExecutionMode() == > This is not an 'eq' condition. We're getting everything from the index. Done -- To view, visit https://asterix-gerrit.ics.uci.edu/2111 To unsubscribe, visit https://asterix-gerrit.ics.uci.edu/settings Gerrit-MessageType: comment Gerrit-Change-Id: I3bbb2b5e1f25e61928d73b866e91c592ce0bf954 Gerrit-PatchSet: 8 Gerrit-Project: asterixdb Gerrit-Branch: master Gerrit-Owner: Ali Alsuliman <[email protected]> Gerrit-Reviewer: Ali Alsuliman <[email protected]> Gerrit-Reviewer: Anon. E. Moose #1000171 Gerrit-Reviewer: Dmitry Lychagin <[email protected]> Gerrit-Reviewer: Jenkins <[email protected]> Gerrit-Reviewer: Taewoo Kim <[email protected]> Gerrit-Reviewer: Till Westmann <[email protected]> Gerrit-HasComments: Yes
