Add Semi Filter to Join Fuzzer (facebookincubator#11473)

Summary:

This changes adds a semi filter to the join filter 10% of the time. Currently it only supports boolean columns. The next steps will be to support integer columns.

Differential Revision: D65629460

velox/exec/fuzzer/JoinFuzzer.cpp

@@ -130,6 +130,17 @@ class JoinFuzzer {
   // Randomly pick a join type to test.
   core::JoinType pickJoinType();
 
+  // Returns an equality join filter between probeKeys and buildKeys and adds a
+  // semi filter 10% of the time. When there is a semi filter, 50% of the time
+  // it is based off a column from the probe side, and the rest of the time, the
+  // build side. A random column is chosen from the probe or build side to
+  // filter on.
+  std::string makeJoinFilter(
+      const std::vector<std::string>& probeKeys,
+      const std::vector<std::string>& buildKeys,
+      const std::vector<RowVectorPtr>& probeInput,
+      const std::vector<RowVectorPtr>& buildInput);
+
   // Makes the query plan with default settings in JoinFuzzer and value inputs
   // for both probe and build sides.
   //
@@ -380,6 +391,38 @@ core::JoinType JoinFuzzer::pickJoinType() {
   return kJoinTypes[idx];
 }
 
+std::string JoinFuzzer::makeJoinFilter(
+    const std::vector<std::string>& probeKeys,
+    const std::vector<std::string>& buildKeys,
+    const std::vector<RowVectorPtr>& probeInput,
+    const std::vector<RowVectorPtr>& buildInput) {
+  const auto numKeys = probeKeys.size();
+  std::string filter;
+  VELOX_CHECK_EQ(numKeys, buildKeys.size());
+  for (auto i = 0; i < numKeys; ++i) {
+    if (i > 0) {
+      filter += " AND ";
+    }
+    filter += fmt::format("{} = {}", probeKeys[i], buildKeys[i]);
+  }
+  // Add a semi filter 10% of the time.
+  if (vectorFuzzer_.coinToss(0.1) && !probeInput.empty() &&
+      !buildInput.empty()) {
+    RowTypePtr rowType = vectorFuzzer_.coinToss(0.5)
+        ? asRowType(probeInput[0]->type())
+        : asRowType(buildInput[0]->type());
+
+    for (int i = 0; i < rowType->size(); i++) {
+      // TODO: Add support for non-boolean types.
+      if (rowType->childAt(i)->isBoolean()) {
+        filter += fmt::format(" AND {} = true", rowType->nameOf(i));
+        break;
+      }
+    }
+  }
+  return filter;
+}
+
 std::vector<TypePtr> JoinFuzzer::generateJoinKeyTypes(int32_t numKeys) {
   std::vector<TypePtr> types;
   types.reserve(numKeys);
@@ -690,6 +733,8 @@ JoinFuzzer::PlanWithSplits JoinFuzzer::makeDefaultPlan(
     const std::vector<RowVectorPtr>& buildInput,
     const std::vector<std::string>& outputColumns) {
   auto planNodeIdGenerator = std::make_shared<core::PlanNodeIdGenerator>();
+  std::string filter =
+      makeJoinFilter(probeKeys, buildKeys, probeInput, buildInput);
   auto plan =
       PlanBuilder(planNodeIdGenerator)
           .values(probeInput)
@@ -773,22 +818,6 @@ std::vector<core::PlanNodePtr> makeSources(
   return sourceNodes;
 }
 
-// Returns an equality join filter between probeKeys and buildKeys.
-std::string makeJoinFilter(
-    const std::vector<std::string>& probeKeys,
-    const std::vector<std::string>& buildKeys) {
-  const auto numKeys = probeKeys.size();
-  std::string filter;
-  VELOX_CHECK_EQ(numKeys, buildKeys.size());
-  for (auto i = 0; i < numKeys; ++i) {
-    if (i > 0) {
-      filter += " AND ";
-    }
-    filter += fmt::format("{} = {}", probeKeys[i], buildKeys[i]);
-  }
-  return filter;
-}
-
 template <typename TNode>
 void addFlippedJoinPlan(
     const core::PlanNodePtr& plan,
@@ -846,8 +875,9 @@ JoinFuzzer::PlanWithSplits JoinFuzzer::makeNestedLoopJoinPlan(
     const std::vector<std::string>& outputColumns,
     bool withFilter) {
   auto planNodeIdGenerator = std::make_shared<core::PlanNodeIdGenerator>();
-  const std::string filter =
-      withFilter ? makeJoinFilter(probeKeys, buildKeys) : "";
+  const std::string filter = withFilter
+      ? makeJoinFilter(probeKeys, buildKeys, probeInput, buildInput)
+      : "";
   return JoinFuzzer::PlanWithSplits{
       PlanBuilder(planNodeIdGenerator)
           .values(probeInput)
@@ -1232,7 +1262,7 @@ JoinFuzzer::PlanWithSplits JoinFuzzer::makeNestedLoopJoinPlanWithTableScan(
   core::PlanNodeId buildScanId;
 
   const std::string filter =
-      withFilter ? makeJoinFilter(probeKeys, buildKeys) : "";
+      withFilter ? makeJoinFilter(probeKeys, buildKeys, {}, {}) : "";
   return JoinFuzzer::PlanWithSplits{
       PlanBuilder(planNodeIdGenerator)
           .tableScan(probeType)