Function TEST_F

bolt/exec/tests/HashJoinTest.cpp:3617–3669

Verify that dynamic filter pushed down from null-aware right semi project join into table scan doesn't filter out nulls.

      {"t0"},
      {
          makeNullableFlatVector<int32_t>({1, std::nullopt, 2}),
      });

  auto build = makeRowVector(
      {"u0"},
      {
          makeNullableFlatVector<int32_t>({1, 2, 3, std::nullopt}),
      });

  std::shared_ptr<TempFilePath> probeFile = TempFilePath::create();
  writeToFile(probeFile->path, {probe});

  std::shared_ptr<TempFilePath> buildFile = TempFilePath::create();
  writeToFile(buildFile->path, {build});

  createDuckDbTable("t", {probe});
  createDuckDbTable("u", {build});

  core::PlanNodeId probeScanId;
  core::PlanNodeId buildScanId;
  auto planNodeIdGenerator = std::make_shared<core::PlanNodeIdGenerator>();
  auto plan = PlanBuilder(planNodeIdGenerator)
                  .tableScan(asRowType(probe->type()))
                  .capturePlanNodeId(probeScanId)
                  .hashJoin(
                      {"t0"},
                      {"u0"},
                      PlanBuilder(planNodeIdGenerator)
                          .tableScan(asRowType(build->type()))
                          .capturePlanNodeId(buildScanId)
                          .planNode(),
                      "",
                      {"u0", "match"},
                      core::JoinType::kRightSemiProject,
                      true /*nullAware*/)
                  .planNode();

  SplitInput splitInput = {
      {probeScanId, {exec::Split(makeHiveConnectorSplit(probeFile->path))}},
      {buildScanId, {exec::Split(makeHiveConnectorSplit(buildFile->path))}},
  };

  HashJoinBuilder(*pool_, duckDbQueryRunner_, driverExecutor_.get())
      .planNode(plan)
      .inputSplits(splitInput)
      .checkSpillStats(false)
      .referenceQuery("SELECT u0, u0 IN (SELECT t0 FROM t) FROM u")
      .run();
}

TEST_F(HashJoinTest, duplicateJoinKeys) {
  auto leftVectors = makeBatches(3, [&](int32_t /*unused*/) {
    return makeRowVector({
        makeNullableFlatVector<int64_t>(
            {1, 2, 2, 3, 3, std::nullopt, 4, 5, 5, 6, 7}),
        makeNullableFlatVector<int64_t>(