Add kRange preceding/following frames in window fuzzer (facebookincub…

…ator#10006)

Summary:
1. Adds support for kRange preceding/following frames in window fuzzer
2. Adds [reference query runner context](https://ibm.box.com/s/9tuk22hfp13imjwq9xelnz1dogsdgdh1) for PrestoSql frame clause for kRange
preceding/following frames.

Resolves facebookincubator#9572 .

Pull Request resolved: facebookincubator#10006

Reviewed By: xiaoxmeng

Differential Revision: D61882004

Pulled By: kagamiori

fbshipit-source-id: 343201fc36e5c3c01779d73ff0888cd0597ba13c

velox/exec/fuzzer/AggregationFuzzerBase.cpp

@@ -241,14 +241,35 @@ std::vector<std::string> AggregationFuzzerBase::generateKeys(
 std::vector<std::string> AggregationFuzzerBase::generateSortingKeys(
     const std::string& prefix,
     std::vector<std::string>& names,
-    std::vector<TypePtr>& types) {
+    std::vector<TypePtr>& types,
+    bool rangeFrame) {
   std::vector<std::string> keys;
-  auto numKeys = boost::random::uniform_int_distribution<uint32_t>(1, 5)(rng_);
+  vector_size_t numKeys;
+  vector_size_t maxDepth;
+  std::vector<TypePtr> sortingKeyTypes = defaultScalarTypes();
+
+  // If frame has k-RANGE bound, only one sorting key should be present, and it
+  // should be a scalar type which supports '+', '-' arithmetic operations.
+  if (rangeFrame) {
+    numKeys = 1;
+    sortingKeyTypes = {
+        TINYINT(),
+        SMALLINT(),
+        INTEGER(),
+        BIGINT(),
+        HUGEINT(),
+        REAL(),
+        DOUBLE()};
+    maxDepth = 0;
+  } else {
+    numKeys = randInt(1, 5);
+    // Pick random, possibly complex, type.
+    maxDepth = 2;
+  }
+
   for (auto i = 0; i < numKeys; ++i) {
     keys.push_back(fmt::format("{}{}", prefix, i));
-
-    // Pick random, possibly complex, type.
-    types.push_back(vectorFuzzer_.randOrderableType(2));
+    types.push_back(vectorFuzzer_.randOrderableType(sortingKeyTypes, maxDepth));
     names.push_back(keys.back());
   }
 

velox/exec/fuzzer/AggregationFuzzerBase.h

@@ -183,11 +183,13 @@ class AggregationFuzzerBase {
       std::vector<TypePtr>& types);
 
   // Similar to generateKeys, but restricts types to orderable types (i.e. no
-  // maps).
+  // maps). For k-RANGE frame bounds, rangeFrame must be set to true so only
+  // one sorting key is generated.
   std::vector<std::string> generateSortingKeys(
       const std::string& prefix,
       std::vector<std::string>& names,
-      std::vector<TypePtr>& types);
+      std::vector<TypePtr>& types,
+      bool rangeFrame = false);
 
   std::pair<CallableSignature, SignatureStats&> pickSignature();
 

velox/exec/fuzzer/PrestoQueryRunner.cpp

@@ -154,6 +154,7 @@ PrestoQueryRunner::PrestoQueryRunner(
       timeout_(timeout) {
   eventBaseThread_.start("PrestoQueryRunner");
   pool_ = aggregatePool()->addLeafChild("leaf");
+  queryRunnerContext_ = std::make_shared<QueryRunnerContext>();
 }
 
 std::optional<std::string> PrestoQueryRunner::toSql(
@@ -352,56 +353,6 @@ std::optional<std::string> PrestoQueryRunner::toSql(
   return sql.str();
 }
 
-namespace {
-
-void appendWindowFrame(
-    const core::WindowNode::Frame& frame,
-    std::stringstream& sql) {
-  // TODO: Add support for k Range Frames by retrieving the original range bound
-  // from WindowNode.
-  switch (frame.type) {
-    case core::WindowNode::WindowType::kRange:
-      sql << " RANGE";
-      break;
-    case core::WindowNode::WindowType::kRows:
-      sql << " ROWS";
-      break;
-    default:
-      VELOX_UNREACHABLE();
-  }
-  sql << " BETWEEN";
-
-  auto appendBound = [&sql](
-                         const core::WindowNode::BoundType& bound,
-                         const core::TypedExprPtr& value) {
-    switch (bound) {
-      case core::WindowNode::BoundType::kUnboundedPreceding:
-        sql << " UNBOUNDED PRECEDING";
-        break;
-      case core::WindowNode::BoundType::kUnboundedFollowing:
-        sql << " UNBOUNDED FOLLOWING";
-        break;
-      case core::WindowNode::BoundType::kCurrentRow:
-        sql << " CURRENT ROW";
-        break;
-      case core::WindowNode::BoundType::kPreceding:
-        sql << " " << value->toString() << " PRECEDING";
-        break;
-      case core::WindowNode::BoundType::kFollowing:
-        sql << " " << value->toString() << " FOLLOWING";
-        break;
-      default:
-        VELOX_UNREACHABLE();
-    }
-  };
-
-  appendBound(frame.startType, frame.startValue);
-  sql << " AND";
-  appendBound(frame.endType, frame.endValue);
-}
-
-} // namespace
-
 std::optional<std::string> PrestoQueryRunner::toSql(
     const std::shared_ptr<const core::WindowNode>& windowNode) {
   if (!isSupportedDwrfType(windowNode->sources()[0]->outputType())) {
@@ -446,7 +397,7 @@ std::optional<std::string> PrestoQueryRunner::toSql(
       }
     }
 
-    appendWindowFrame(functions[i].frame, sql);
+    sql << " " << queryRunnerContext_->windowFrames_.at(windowNode->id()).at(i);
     sql << ")";
   }
 

velox/exec/fuzzer/PrestoQueryRunner.h

@@ -31,6 +31,11 @@ T extractSingleValue(const std::vector<RowVectorPtr>& data) {
   return simpleVector->valueAt(0);
 }
 
+class QueryRunnerContext {
+ public:
+  std::unordered_map<core::PlanNodeId, std::vector<std::string>> windowFrames_;
+};
+
 class PrestoQueryRunner : public velox::exec::test::ReferenceQueryRunner {
  public:
   /// @param coordinatorUri Presto REST API endpoint, e.g. http://127.0.0.1:8080
@@ -99,6 +104,10 @@ class PrestoQueryRunner : public velox::exec::test::ReferenceQueryRunner {
       const std::vector<RowVectorPtr>& buildInput,
       const RowTypePtr& resultType) override;
 
+  std::shared_ptr<QueryRunnerContext> queryRunnerContext() {
+    return queryRunnerContext_;
+  }
+
  private:
   memory::MemoryPool* pool() {
     return pool_.get();
@@ -141,6 +150,7 @@ class PrestoQueryRunner : public velox::exec::test::ReferenceQueryRunner {
   const std::chrono::milliseconds timeout_;
   folly::EventBaseThread eventBaseThread_{false};
   std::shared_ptr<memory::MemoryPool> pool_;
+  std::shared_ptr<QueryRunnerContext> queryRunnerContext_;
 };
 
 } // namespace facebook::velox::exec::test