Merge branch 'master' into T-fix-orderby-spike

.github/workflows/daily_benchmark.yaml

@@ -75,6 +75,8 @@ jobs:
 
           ./benchmark.py vendor-native --num-workers-for-benchmark 12 --export-results cartesian.json cartesian
 
+          ./benchmark.py vendor-native --num-workers-for-benchmark 12 --export-results order_by.json order_by
+
       - name: Upload mgbench results
         run: |
           cd tools/bench-graph-client
@@ -104,3 +106,9 @@ jobs:
                     --github-run-id "${{ github.run_id }}" \
                     --github-run-number "${{ github.run_number }}" \
                     --head-branch-name "${{ env.BRANCH_NAME }}"
+
+          ./main.py --benchmark-name "order_by" \
+                    --benchmark-results "../../tests/mgbench/order_by.json" \
+                    --github-run-id "${{ github.run_id }}" \
+                    --github-run-number "${{ github.run_number }}" \
+                    --head-branch-name "${{ env.BRANCH_NAME }}"

src/query/interpreter.cpp

@@ -238,6 +238,16 @@ std::optional<std::string> GetOptionalStringValue(query::Expression *expression,
   return {};
 };
 
+bool IsOrderByQuery(const std::vector<memgraph::query::Clause *> &clauses) {
+  for (const auto &clause : clauses) {
+    if (clause->GetTypeInfo() == Return::kType) {
+      auto *return_clause = utils::Downcast<Return>(clause);
+      return !return_clause->body_.order_by.empty();
+    }
+  }
+  return false;
+}
+
 bool IsAllShortestPathsQuery(const std::vector<memgraph::query::Clause *> &clauses) {
   for (const auto &clause : clauses) {
     if (clause->GetTypeInfo() != Match::kType) {
@@ -1591,8 +1601,8 @@ PreparedQuery PrepareCypherQuery(ParsedQuery parsed_query, std::map<std::string,
   }
 
   // If this is LOAD CSV query, use PoolResource without MonotonicMemoryResource as we want to reuse allocated memory
-  auto use_monotonic_memory =
-      !contains_csv && !IsCallBatchedProcedureQuery(clauses) && !IsAllShortestPathsQuery(clauses);
+  auto use_monotonic_memory = !contains_csv && !IsCallBatchedProcedureQuery(clauses) &&
+                              !IsAllShortestPathsQuery(clauses) && !IsOrderByQuery(clauses);
 
   MG_ASSERT(current_db.execution_db_accessor_, "Cypher query expects a current DB transaction");
   auto *dba =
@@ -1754,8 +1764,8 @@ PreparedQuery PrepareProfileQuery(ParsedQuery parsed_query, bool in_explicit_tra
 
   // If this is LOAD CSV, BatchedProcedure or AllShortest query, use PoolResource without MonotonicMemoryResource as we
   // want to reuse allocated memory
-  auto use_monotonic_memory =
-      !contains_csv && !IsCallBatchedProcedureQuery(clauses) && !IsAllShortestPathsQuery(clauses);
+  auto use_monotonic_memory = !contains_csv && !IsCallBatchedProcedureQuery(clauses) &&
+                              !IsAllShortestPathsQuery(clauses) && !IsOrderByQuery(clauses);
 
   MG_ASSERT(cypher_query, "Cypher grammar should not allow other queries in PROFILE");
   EvaluationContext evaluation_context;
@@ -3716,7 +3726,8 @@ Interpreter::PrepareResult Interpreter::Prepare(const std::string &query_string,
       auto const &clauses = cypher_query->single_query_->clauses_;
       bool hasAllShortestPaths = IsAllShortestPathsQuery(clauses);
       // Using PoolResource without MonotonicMemoryResouce for LOAD CSV reduces memory usage.
-      bool usePool = hasAllShortestPaths || IsCallBatchedProcedureQuery(clauses) || IsLoadCsvQuery(clauses);
+      bool usePool = hasAllShortestPaths || IsCallBatchedProcedureQuery(clauses) || IsLoadCsvQuery(clauses) ||
+                     IsOrderByQuery(clauses);
       return {usePool, hasAllShortestPaths};
     }();  // IILE
 

tests/mgbench/workloads/order_by.py

@@ -0,0 +1,46 @@
+# Copyright 2023 Memgraph Ltd.
+#
+# Use of this software is governed by the Business Source License
+# included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
+# License, and you may not use this file except in compliance with the Business Source License.
+#
+# As of the Change Date specified in that file, in accordance with
+# the Business Source License, use of this software will be governed
+# by the Apache License, Version 2.0, included in the file
+# licenses/APL.txt.
+
+from workloads.base import Workload
+
+
+class OrderBy(Workload):
+    NAME = "order_by"
+    CARDINALITY = 10000
+
+    def indexes_generator(self):
+        return [
+            ("CREATE INDEX ON :Node;", {}),
+            ("CREATE INDEX ON :Node(prop1);", {}),
+            ("CREATE INDEX ON :Node(prop2);", {}),
+            ("CREATE INDEX ON :Node(prop3);", {}),
+        ]
+
+    def dataset_generator(self):
+        queries = []
+        for i in range(0, OrderBy.CARDINALITY):
+            queries.append(
+                (
+                    """CREATE
+(:Node {prop1: $id, prop2: $id, prop3: $id
+});
+""",
+                    {"id": i},
+                )
+            )
+
+        return queries
+
+    def benchmark__test__order_by(self):
+        return (
+            "MATCH (n:Node) RETURN n ORDER BY n.prop1",
+            {},
+        )