push filter down index scan (#5019)

* push filter down index scan * fix bug Co-authored-by: Sophie <[email protected]>
vesoft-inc · Dec 8, 2022 · 55212e2 · 55212e2
1 parent 5057eeb
commit 55212e2
Show file tree

Hide file tree

Showing 3 changed files with 50 additions and 9 deletions.
diff --git a/src/graph/optimizer/rule/PushLimitDownScanAppendVerticesRule.cpp b/src/graph/optimizer/rule/PushLimitDownScanAppendVerticesRule.cpp
@@ -11,6 +11,8 @@
 #include "graph/planner/plan/Query.h"
 
 using nebula::graph::AppendVertices;
+using nebula::graph::Explore;
+using nebula::graph::IndexScan;
 using nebula::graph::Limit;
 using nebula::graph::PlanNode;
 using nebula::graph::QueryContext;
@@ -19,7 +21,8 @@ using nebula::graph::ScanVertices;
 namespace nebula {
 namespace opt {
 
-// Limit->AppendVertices->ScanVertices ==> Limit->AppendVertices->ScanVertices(Limit)
+// Limit->AppendVertices->ScanVertices/IndexScan ==>
+// Limit->AppendVertices->ScanVertices/IndexScan(Limit)
 
 std::unique_ptr<OptRule> PushLimitDownScanAppendVerticesRule::kInstance =
     std::unique_ptr<PushLimitDownScanAppendVerticesRule>(new PushLimitDownScanAppendVerticesRule());
@@ -32,7 +35,8 @@ const Pattern &PushLimitDownScanAppendVerticesRule::pattern() const {
   static Pattern pattern =
       Pattern::create(graph::PlanNode::Kind::kLimit,
                       {Pattern::create(graph::PlanNode::Kind::kAppendVertices,
-                                       {Pattern::create(graph::PlanNode::Kind::kScanVertices)})});
+                                       {Pattern::create({graph::PlanNode::Kind::kScanVertices,
+                                                         graph::PlanNode::Kind::kIndexScan})})});
   return pattern;
 }
 
@@ -53,8 +57,22 @@ bool PushLimitDownScanAppendVerticesRule::match(OptContext *ctx,
   }
   auto *filter = av->filter();
   auto *vFilter = av->vFilter();
-  // Limit can't push over filter operation
-  return filter == nullptr && vFilter == nullptr;
+
+  if (vFilter != nullptr) return false;
+  if (filter == nullptr) return true;
+  auto scanNode = matched.planNode({0, 0, 0});
+  if (scanNode->kind() == PlanNode::Kind::kScanVertices) return false;
+  // If the scanNode is kIndexScan, and the filter looks like `player._tag IS NOT EMPTY`,
+  // the limit could also be pushed down
+  if (filter->kind() != Expression::Kind::kIsNotEmpty) {
+    return false;
+  }
+  auto *isNotEmpty = static_cast<const UnaryExpression *>(filter);
+  auto *operand = isNotEmpty->operand();
+  if (operand->kind() != Expression::Kind::kTagProperty) {
+    return false;
+  }
+  return true;
 }
 
 StatusOr<OptRule::TransformResult> PushLimitDownScanAppendVerticesRule::transform(
@@ -65,7 +83,7 @@ StatusOr<OptRule::TransformResult> PushLimitDownScanAppendVerticesRule::transfor
 
   const auto limit = static_cast<const Limit *>(limitGroupNode->node());
   const auto appendVertices = static_cast<const AppendVertices *>(appendVerticesGroupNode->node());
-  const auto scanVertices = static_cast<const ScanVertices *>(scanVerticesGroupNode->node());
+  const auto scanVertices = static_cast<const Explore *>(scanVerticesGroupNode->node());
 
   int64_t limitRows = limit->offset() + limit->count();
   if (scanVertices->limit() >= 0 && limitRows >= scanVertices->limit()) {
@@ -80,7 +98,7 @@ StatusOr<OptRule::TransformResult> PushLimitDownScanAppendVerticesRule::transfor
   auto newAppendVerticesGroup = OptGroup::create(octx);
   auto newAppendVerticesGroupNode = newAppendVerticesGroup->makeGroupNode(newAppendVertices);
 
-  auto newScanVertices = static_cast<ScanVertices *>(scanVertices->clone());
+  auto newScanVertices = static_cast<Explore *>(scanVertices->clone());
   newScanVertices->setLimit(limitRows);
   auto newScanVerticesGroup = OptGroup::create(octx);
   auto newScanVerticesGroupNode = newScanVerticesGroup->makeGroupNode(newScanVertices);

diff --git a/src/graph/optimizer/rule/PushLimitDownScanAppendVerticesRule.h b/src/graph/optimizer/rule/PushLimitDownScanAppendVerticesRule.h
@@ -30,9 +30,9 @@ namespace opt {
 // |   AppendVertices  |
 // +---------+---------+
 //           |
-// +---------+---------+
-// |    ScanVertices   |
-// +---------+---------+
+// +---------+-------------------+
+// |    ScanVertices/IndexScan   |
+// +---------+-------------------+
 //
 //  After:
 //
@@ -47,6 +47,7 @@ namespace opt {
 //           |
 // +---------+---------+
 // |    ScanVertices   |
+// |    IndexScan      |
 // |     (limit=3)     |
 // +---------+---------+
 

diff --git a/tests/tck/features/match/Base.feature b/tests/tck/features/match/Base.feature
@@ -14,6 +14,22 @@ Feature: Basic match
     Then the result should be, in any order, with relax comparison:
       | v            |
       | ("Yao Ming") |
+    When profiling query:
+      """
+      MATCH (v:player) RETURN v.player.age LIMIT 3
+      """
+    Then the result should be, in any order, with relax comparison:
+      | v.player.age |
+      | /\d+/        |
+      | /\d+/        |
+      | /\d+/        |
+    Then the execution plan should be:
+      | id | name           | dependencies | operator info                 |
+      | 5  | Project        | 4            |                               |
+      | 4  | Limit          | 3            | {"offset": "0", "count": "3"} |
+      | 3  | AppendVertices | 2            |                               |
+      | 2  | IndexScan      | 1            | {"limit": "3"}                |
+      | 1  | Start          |              |                               |
     When executing query:
       """
       MATCH (v:player) WHERE v.player.age < 0 RETURN v
@@ -629,6 +645,12 @@ Feature: Basic match
       MATCH (v:player:bachelor) RETURN v
       """
     Then a ExecutionError should be raised at runtime: Scan vertices or edges need to specify a limit number, or limit number can not push down.
+    # TODO(jie): Optimize this case
+    When executing query:
+      """
+      MATCH (v) WHERE v.player.age == 18 RETURN v LIMIT 3
+      """
+    Then a ExecutionError should be raised at runtime: Scan vertices or edges need to specify a limit number, or limit number can not push down.
     When executing query:
       """
       MATCH (v:player{age:23}:bachelor) RETURN v