[improve](nereids) support pushdown count when scan project without a…

…ny slot ref (apache#35162) select 66 from baseall_dup; could use pushAggOp=COUNT, so no need to scan real data
xyfsjq · May 27, 2024 · b66f9ff · b66f9ff
1 parent 717722e
commit b66f9ff
Show file tree

Hide file tree

Showing 5 changed files with 55 additions and 5 deletions.
diff --git a/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/RuleType.java b/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/RuleType.java
@@ -432,6 +432,7 @@ public enum RuleType {
     STORAGE_LAYER_AGGREGATE_WITH_PROJECT(RuleTypeClass.IMPLEMENTATION),
     STORAGE_LAYER_AGGREGATE_WITHOUT_PROJECT_FOR_FILE_SCAN(RuleTypeClass.IMPLEMENTATION),
     STORAGE_LAYER_AGGREGATE_WITH_PROJECT_FOR_FILE_SCAN(RuleTypeClass.IMPLEMENTATION),
+    STORAGE_LAYER_WITH_PROJECT_NO_SLOT_REF(RuleTypeClass.IMPLEMENTATION),
     STORAGE_LAYER_AGGREGATE_MINMAX_ON_UNIQUE(RuleTypeClass.IMPLEMENTATION),
     STORAGE_LAYER_AGGREGATE_MINMAX_ON_UNIQUE_WITHOUT_PROJECT(RuleTypeClass.IMPLEMENTATION),
     COUNT_ON_INDEX(RuleTypeClass.IMPLEMENTATION),

diff --git a/...core/src/main/java/org/apache/doris/nereids/rules/implementation/AggregateStrategies.java b/...core/src/main/java/org/apache/doris/nereids/rules/implementation/AggregateStrategies.java
@@ -215,6 +215,16 @@ public List<Rule> buildRules() {
                 .when(agg -> agg.isNormalized() && enablePushDownNoGroupAgg())
                 .thenApply(ctx -> storageLayerAggregate(ctx.root, null, ctx.root.child(), ctx.cascadesContext))
             ),
+            RuleType.STORAGE_LAYER_WITH_PROJECT_NO_SLOT_REF.build(
+                    logicalProject(
+                            logicalOlapScan()
+                    )
+                    .thenApply(ctx -> {
+                        LogicalProject<LogicalOlapScan> project = ctx.root;
+                        LogicalOlapScan olapScan = project.child();
+                        return pushDownCountWithoutSlotRef(project, olapScan, ctx.cascadesContext);
+                    })
+            ),
             RuleType.STORAGE_LAYER_AGGREGATE_WITH_PROJECT.build(
                 logicalAggregate(
                     logicalProject(
@@ -306,6 +316,37 @@ && couldConvertToMulti(agg))
         );
     }
 
+    /*
+     *  select 66 from baseall_dup; could use pushAggOp=COUNT to not scan real data.
+     */
+    private LogicalProject<? extends Plan> pushDownCountWithoutSlotRef(
+            LogicalProject<? extends Plan> project,
+            LogicalOlapScan logicalScan,
+            CascadesContext cascadesContext) {
+        final LogicalProject<? extends Plan> canNotPush = project;
+        if (!enablePushDownNoGroupAgg()) {
+            return canNotPush;
+        }
+        if (logicalScan != null) {
+            KeysType keysType = logicalScan.getTable().getKeysType();
+            if (keysType != KeysType.DUP_KEYS) {
+                return canNotPush;
+            }
+        }
+        for (Expression e : project.getProjects()) {
+            if (e.anyMatch(SlotReference.class::isInstance)) {
+                return canNotPush;
+            }
+        }
+        PhysicalOlapScan physicalOlapScan
+                = (PhysicalOlapScan) new LogicalOlapScanToPhysicalOlapScan()
+                .build()
+                .transform(logicalScan, cascadesContext)
+                .get(0);
+        return project.withChildren(ImmutableList.of(new PhysicalStorageLayerAggregate(
+                physicalOlapScan, PushDownAggOp.COUNT)));
+    }
+
     private boolean enablePushDownMinMaxOnUnique() {
         ConnectContext connectContext = ConnectContext.get();
         return connectContext != null && connectContext.getSessionVariable().isEnablePushDownMinMaxOnUnique();

diff --git a/...on-test/data/nereids_rules_p0/eliminate_distinct_constant/eliminate_distinct_constant.out b/...on-test/data/nereids_rules_p0/eliminate_distinct_constant/eliminate_distinct_constant.out
@@ -4,28 +4,28 @@ PhysicalResultSink
 --PhysicalLimit[GLOBAL]
 ----PhysicalLimit[LOCAL]
 ------PhysicalProject
---------PhysicalOlapScan[t1]
+--------PhysicalStorageLayerAggregate[t1]
 
 -- !basic_2 --
 PhysicalResultSink
 --PhysicalLimit[GLOBAL]
 ----PhysicalLimit[LOCAL]
 ------PhysicalProject
---------PhysicalOlapScan[t2]
+--------PhysicalStorageLayerAggregate[t2]
 
 -- !basic_3 --
 PhysicalResultSink
 --PhysicalLimit[GLOBAL]
 ----PhysicalLimit[LOCAL]
 ------PhysicalProject
---------PhysicalOlapScan[t3]
+--------PhysicalStorageLayerAggregate[t3]
 
 -- !basic_4 --
 PhysicalResultSink
 --PhysicalLimit[GLOBAL]
 ----PhysicalLimit[LOCAL]
 ------PhysicalProject
---------PhysicalOlapScan[t4]
+--------PhysicalStorageLayerAggregate[t4]
 
 -- !basic_1 --
 1	2	3

diff --git a/regression-test/data/nereids_rules_p0/filter_push_down/push_down_filter_other_condition.out b/regression-test/data/nereids_rules_p0/filter_push_down/push_down_filter_other_condition.out
@@ -241,5 +241,5 @@ PhysicalResultSink
 ------PhysicalOlapScan[t1]
 ----PhysicalLimit[GLOBAL]
 ------PhysicalLimit[LOCAL]
---------PhysicalOlapScan[t2]
+--------PhysicalStorageLayerAggregate[t2]
 
diff --git a/regression-test/suites/nereids_p0/explain/test_pushdown_explain.groovy b/regression-test/suites/nereids_p0/explain/test_pushdown_explain.groovy
@@ -65,6 +65,14 @@ suite("test_pushdown_explain") {
         sql("select count(cast(lo_orderkey as bigint)) from test_lineorder;")
         contains "pushAggOp=COUNT"
     }
+    explain {
+        sql("select 66 from test_lineorder;")
+        contains "pushAggOp=COUNT"
+    }
+    explain {
+        sql("select lo_orderkey from test_lineorder;")
+        contains "pushAggOp=NONE"
+    }
 
     sql "DROP TABLE IF EXISTS table_unique0"
     sql """