Search in sources :

Example 1 with FULL

use of io.trino.sql.planner.plan.JoinNode.Type.FULL in project trino by trinodb.

the class PushJoinIntoTableScan method apply.

@Override
public Result apply(JoinNode joinNode, Captures captures, Context context) {
    if (joinNode.isCrossJoin()) {
        return Result.empty();
    }
    TableScanNode left = captures.get(LEFT_TABLE_SCAN);
    TableScanNode right = captures.get(RIGHT_TABLE_SCAN);
    verify(!left.isUpdateTarget() && !right.isUpdateTarget(), "Unexpected Join over for-update table scan");
    Expression effectiveFilter = getEffectiveFilter(joinNode);
    FilterSplitResult filterSplitResult = splitFilter(effectiveFilter, left.getOutputSymbols(), right.getOutputSymbols(), context);
    if (!filterSplitResult.getRemainingFilter().equals(BooleanLiteral.TRUE_LITERAL)) {
        // TODO add extra filter node above join
        return Result.empty();
    }
    if (left.getEnforcedConstraint().isNone() || right.getEnforcedConstraint().isNone()) {
        // enforced constraint harder below.
        return Result.empty();
    }
    Map<String, ColumnHandle> leftAssignments = left.getAssignments().entrySet().stream().collect(toImmutableMap(entry -> entry.getKey().getName(), Map.Entry::getValue));
    Map<String, ColumnHandle> rightAssignments = right.getAssignments().entrySet().stream().collect(toImmutableMap(entry -> entry.getKey().getName(), Map.Entry::getValue));
    /*
         * We are (lazily) computing estimated statistics for join node and left and right table
         * and passing those to connector via applyJoin.
         *
         * There are a couple reasons for this approach:
         * - the engine knows how to estimate join and connector may not
         * - the engine may have cached stats for the table scans (within context.getStatsProvider()), so can be able to provide information more inexpensively
         * - in the future, the engine may be able to provide stats for table scan even in case when connector no longer can (see https://github.com/trinodb/trino/issues/6998)
         * - the pushdown feasibility assessment logic may be different (or configured differently) for different connectors/catalogs.
         */
    JoinStatistics joinStatistics = getJoinStatistics(joinNode, left, right, context);
    Optional<JoinApplicationResult<TableHandle>> joinApplicationResult = metadata.applyJoin(context.getSession(), getJoinType(joinNode), left.getTable(), right.getTable(), filterSplitResult.getPushableConditions(), // TODO we could pass only subset of assignments here, those which are needed to resolve filterSplitResult.getPushableConditions
    leftAssignments, rightAssignments, joinStatistics);
    if (joinApplicationResult.isEmpty()) {
        return Result.empty();
    }
    TableHandle handle = joinApplicationResult.get().getTableHandle();
    Map<ColumnHandle, ColumnHandle> leftColumnHandlesMapping = joinApplicationResult.get().getLeftColumnHandles();
    Map<ColumnHandle, ColumnHandle> rightColumnHandlesMapping = joinApplicationResult.get().getRightColumnHandles();
    ImmutableMap.Builder<Symbol, ColumnHandle> assignmentsBuilder = ImmutableMap.builder();
    assignmentsBuilder.putAll(left.getAssignments().entrySet().stream().collect(toImmutableMap(Map.Entry::getKey, entry -> leftColumnHandlesMapping.get(entry.getValue()))));
    assignmentsBuilder.putAll(right.getAssignments().entrySet().stream().collect(toImmutableMap(Map.Entry::getKey, entry -> rightColumnHandlesMapping.get(entry.getValue()))));
    Map<Symbol, ColumnHandle> assignments = assignmentsBuilder.buildOrThrow();
    // convert enforced constraint
    JoinNode.Type joinType = joinNode.getType();
    TupleDomain<ColumnHandle> leftConstraint = deriveConstraint(left.getEnforcedConstraint(), leftColumnHandlesMapping, joinType == RIGHT || joinType == FULL);
    TupleDomain<ColumnHandle> rightConstraint = deriveConstraint(right.getEnforcedConstraint(), rightColumnHandlesMapping, joinType == LEFT || joinType == FULL);
    TupleDomain<ColumnHandle> newEnforcedConstraint = TupleDomain.withColumnDomains(ImmutableMap.<ColumnHandle, Domain>builder().putAll(leftConstraint.getDomains().orElseThrow()).putAll(rightConstraint.getDomains().orElseThrow()).buildOrThrow());
    return Result.ofPlanNode(new ProjectNode(context.getIdAllocator().getNextId(), new TableScanNode(joinNode.getId(), handle, ImmutableList.copyOf(assignments.keySet()), assignments, newEnforcedConstraint, deriveTableStatisticsForPushdown(context.getStatsProvider(), context.getSession(), joinApplicationResult.get().isPrecalculateStatistics(), joinNode), false, Optional.empty()), Assignments.identity(joinNode.getOutputSymbols())));
}
Also used : PlanNode(io.trino.sql.planner.plan.PlanNode) LEFT(io.trino.sql.planner.plan.JoinNode.Type.LEFT) BooleanLiteral(io.trino.sql.tree.BooleanLiteral) Map(java.util.Map) JoinNode(io.trino.sql.planner.plan.JoinNode) TableScanNode(io.trino.sql.planner.plan.TableScanNode) PlanNodeStatsEstimate(io.trino.cost.PlanNodeStatsEstimate) Rules.deriveTableStatisticsForPushdown(io.trino.sql.planner.iterative.rule.Rules.deriveTableStatisticsForPushdown) ImmutableSet(com.google.common.collect.ImmutableSet) ImmutableMap(com.google.common.collect.ImmutableMap) Domain(io.trino.spi.predicate.Domain) ImmutableList.toImmutableList(com.google.common.collect.ImmutableList.toImmutableList) Assignments(io.trino.sql.planner.plan.Assignments) Set(java.util.Set) BasicRelationStatistics(io.trino.spi.connector.BasicRelationStatistics) Patterns.tableScan(io.trino.sql.planner.plan.Patterns.tableScan) ComparisonExpression(io.trino.sql.tree.ComparisonExpression) List(java.util.List) ImmutableMap.toImmutableMap(com.google.common.collect.ImmutableMap.toImmutableMap) Pattern(io.trino.matching.Pattern) SystemSessionProperties.isAllowPushdownIntoConnectors(io.trino.SystemSessionProperties.isAllowPushdownIntoConnectors) RIGHT(io.trino.sql.planner.plan.JoinNode.Type.RIGHT) SymbolReference(io.trino.sql.tree.SymbolReference) Optional(java.util.Optional) Expression(io.trino.sql.tree.Expression) ExpressionUtils.extractConjuncts(io.trino.sql.ExpressionUtils.extractConjuncts) Session(io.trino.Session) JoinCondition(io.trino.spi.connector.JoinCondition) Patterns(io.trino.sql.planner.plan.Patterns) Variable(io.trino.spi.expression.Variable) Capture.newCapture(io.trino.matching.Capture.newCapture) JoinStatistics(io.trino.spi.connector.JoinStatistics) ImmutableList(com.google.common.collect.ImmutableList) JoinType(io.trino.spi.connector.JoinType) Verify.verify(com.google.common.base.Verify.verify) Objects.requireNonNull(java.util.Objects.requireNonNull) ColumnHandle(io.trino.spi.connector.ColumnHandle) Rule(io.trino.sql.planner.iterative.Rule) Join.left(io.trino.sql.planner.plan.Patterns.Join.left) ProjectNode(io.trino.sql.planner.plan.ProjectNode) ExpressionUtils(io.trino.sql.ExpressionUtils) Symbol(io.trino.sql.planner.Symbol) FULL(io.trino.sql.planner.plan.JoinNode.Type.FULL) TupleDomain(io.trino.spi.predicate.TupleDomain) Capture(io.trino.matching.Capture) JoinApplicationResult(io.trino.spi.connector.JoinApplicationResult) TableHandle(io.trino.metadata.TableHandle) Join.right(io.trino.sql.planner.plan.Patterns.Join.right) ExpressionUtils.and(io.trino.sql.ExpressionUtils.and) Double.isNaN(java.lang.Double.isNaN) Captures(io.trino.matching.Captures) Metadata(io.trino.metadata.Metadata) TypeProvider(io.trino.sql.planner.TypeProvider) Domain.onlyNull(io.trino.spi.predicate.Domain.onlyNull) ColumnHandle(io.trino.spi.connector.ColumnHandle) Symbol(io.trino.sql.planner.Symbol) JoinNode(io.trino.sql.planner.plan.JoinNode) JoinApplicationResult(io.trino.spi.connector.JoinApplicationResult) ImmutableMap(com.google.common.collect.ImmutableMap) ImmutableMap.toImmutableMap(com.google.common.collect.ImmutableMap.toImmutableMap) JoinStatistics(io.trino.spi.connector.JoinStatistics) TableScanNode(io.trino.sql.planner.plan.TableScanNode) ComparisonExpression(io.trino.sql.tree.ComparisonExpression) Expression(io.trino.sql.tree.Expression) TableHandle(io.trino.metadata.TableHandle) ProjectNode(io.trino.sql.planner.plan.ProjectNode) Domain(io.trino.spi.predicate.Domain) TupleDomain(io.trino.spi.predicate.TupleDomain) Map(java.util.Map) ImmutableMap(com.google.common.collect.ImmutableMap) ImmutableMap.toImmutableMap(com.google.common.collect.ImmutableMap.toImmutableMap)

Example 2 with FULL

use of io.trino.sql.planner.plan.JoinNode.Type.FULL in project trino by trinodb.

the class TestDetermineJoinDistributionType method testPartitionFullOuterJoin.

@Test
public void testPartitionFullOuterJoin() {
    int aRows = 10_000;
    int bRows = 10;
    assertDetermineJoinDistributionType().setSystemProperty(JOIN_DISTRIBUTION_TYPE, JoinDistributionType.AUTOMATIC.name()).overrideStats("valuesA", PlanNodeStatsEstimate.builder().setOutputRowCount(aRows).addSymbolStatistics(ImmutableMap.of(new Symbol("A1"), new SymbolStatsEstimate(0, 100, 0, 640000, 100))).build()).overrideStats("valuesB", PlanNodeStatsEstimate.builder().setOutputRowCount(bRows).addSymbolStatistics(ImmutableMap.of(new Symbol("B1"), new SymbolStatsEstimate(0, 100, 0, 640000, 100))).build()).on(p -> p.join(FULL, p.values(new PlanNodeId("valuesA"), aRows, p.symbol("A1", BIGINT)), p.values(new PlanNodeId("valuesB"), bRows, p.symbol("B1", BIGINT)), ImmutableList.of(new JoinNode.EquiJoinClause(p.symbol("A1", BIGINT), p.symbol("B1", BIGINT))), ImmutableList.of(p.symbol("A1", BIGINT)), ImmutableList.of(p.symbol("B1", BIGINT)), Optional.empty())).matches(join(FULL, ImmutableList.of(equiJoinClause("A1", "B1")), Optional.empty(), Optional.of(PARTITIONED), values(ImmutableMap.of("A1", 0)), values(ImmutableMap.of("B1", 0))));
}
Also used : PARTITIONED(io.trino.sql.planner.plan.JoinNode.DistributionType.PARTITIONED) INNER(io.trino.sql.planner.plan.JoinNode.Type.INNER) VarcharType.createUnboundedVarcharType(io.trino.spi.type.VarcharType.createUnboundedVarcharType) Assert.assertEquals(org.testng.Assert.assertEquals) Test(org.testng.annotations.Test) PlanMatchPattern.filter(io.trino.sql.planner.assertions.PlanMatchPattern.filter) REPLICATED(io.trino.sql.planner.plan.JoinNode.DistributionType.REPLICATED) Lookup.noLookup(io.trino.sql.planner.iterative.Lookup.noLookup) VarcharType(io.trino.spi.type.VarcharType) LEFT(io.trino.sql.planner.plan.JoinNode.Type.LEFT) RuleAssert(io.trino.sql.planner.iterative.rule.test.RuleAssert) Type(io.trino.sql.planner.plan.JoinNode.Type) PlanBuilder.expressions(io.trino.sql.planner.iterative.rule.test.PlanBuilder.expressions) ImmutableList(com.google.common.collect.ImmutableList) NaN(java.lang.Double.NaN) DistributionType(io.trino.sql.planner.plan.JoinNode.DistributionType) PlanNodeId(io.trino.sql.planner.plan.PlanNodeId) PlanBuilder(io.trino.sql.planner.iterative.rule.test.PlanBuilder) PlanMatchPattern.equiJoinClause(io.trino.sql.planner.assertions.PlanMatchPattern.equiJoinClause) JoinNode(io.trino.sql.planner.plan.JoinNode) TableScanNode(io.trino.sql.planner.plan.TableScanNode) PlanMatchPattern.join(io.trino.sql.planner.assertions.PlanMatchPattern.join) PlanNodeStatsEstimate(io.trino.cost.PlanNodeStatsEstimate) JOIN_MAX_BROADCAST_TABLE_SIZE(io.trino.SystemSessionProperties.JOIN_MAX_BROADCAST_TABLE_SIZE) TaskCountEstimator(io.trino.cost.TaskCountEstimator) Symbol(io.trino.sql.planner.Symbol) AfterClass(org.testng.annotations.AfterClass) SymbolStatsEstimate(io.trino.cost.SymbolStatsEstimate) RuleTester.defaultRuleTester(io.trino.sql.planner.iterative.rule.test.RuleTester.defaultRuleTester) ImmutableMap(com.google.common.collect.ImmutableMap) BeforeClass(org.testng.annotations.BeforeClass) FULL(io.trino.sql.planner.plan.JoinNode.Type.FULL) RuleTester(io.trino.sql.planner.iterative.rule.test.RuleTester) PlanMatchPattern.values(io.trino.sql.planner.assertions.PlanMatchPattern.values) TRUE_LITERAL(io.trino.sql.tree.BooleanLiteral.TRUE_LITERAL) JoinDistributionType(io.trino.sql.planner.OptimizerConfig.JoinDistributionType) DetermineJoinDistributionType.getSourceTablesSizeInBytes(io.trino.sql.planner.iterative.rule.DetermineJoinDistributionType.getSourceTablesSizeInBytes) CostComparator(io.trino.cost.CostComparator) PlanMatchPattern.enforceSingleRow(io.trino.sql.planner.assertions.PlanMatchPattern.enforceSingleRow) BIGINT(io.trino.spi.type.BigintType.BIGINT) RIGHT(io.trino.sql.planner.plan.JoinNode.Type.RIGHT) JOIN_DISTRIBUTION_TYPE(io.trino.SystemSessionProperties.JOIN_DISTRIBUTION_TYPE) ImmutableListMultimap(com.google.common.collect.ImmutableListMultimap) Optional(java.util.Optional) ValuesNode(io.trino.sql.planner.plan.ValuesNode) TestingColumnHandle(io.trino.testing.TestingMetadata.TestingColumnHandle) PlanBuilder.expression(io.trino.sql.planner.iterative.rule.test.PlanBuilder.expression) PlanNodeIdAllocator(io.trino.sql.planner.PlanNodeIdAllocator) PlanNodeId(io.trino.sql.planner.plan.PlanNodeId) Symbol(io.trino.sql.planner.Symbol) SymbolStatsEstimate(io.trino.cost.SymbolStatsEstimate) Test(org.testng.annotations.Test)

Example 3 with FULL

use of io.trino.sql.planner.plan.JoinNode.Type.FULL in project trino by trinodb.

the class TestReplaceRedundantJoinWithProject method testReplaceFULLJoin.

@Test
public void testReplaceFULLJoin() {
    tester().assertThat(new ReplaceRedundantJoinWithProject()).on(p -> p.join(FULL, p.values(10, p.symbol("a")), p.values(0, p.symbol("b")))).matches(project(ImmutableMap.of("a", expression("a"), "b", expression("CAST(null AS bigint)")), values(ImmutableList.of("a"), nCopies(10, ImmutableList.of(new NullLiteral())))));
    tester().assertThat(new ReplaceRedundantJoinWithProject()).on(p -> p.join(FULL, p.values(0, p.symbol("a")), p.values(10, p.symbol("b")))).matches(project(ImmutableMap.of("a", expression("CAST(null AS bigint)"), "b", expression("b")), values(ImmutableList.of("b"), nCopies(10, ImmutableList.of(new NullLiteral())))));
}
Also used : PlanMatchPattern.expression(io.trino.sql.planner.assertions.PlanMatchPattern.expression) ImmutableMap(com.google.common.collect.ImmutableMap) BaseRuleTest(io.trino.sql.planner.iterative.rule.test.BaseRuleTest) INNER(io.trino.sql.planner.plan.JoinNode.Type.INNER) Collections.nCopies(java.util.Collections.nCopies) FULL(io.trino.sql.planner.plan.JoinNode.Type.FULL) Test(org.testng.annotations.Test) PlanMatchPattern.values(io.trino.sql.planner.assertions.PlanMatchPattern.values) LEFT(io.trino.sql.planner.plan.JoinNode.Type.LEFT) ImmutableList(com.google.common.collect.ImmutableList) PlanMatchPattern.project(io.trino.sql.planner.assertions.PlanMatchPattern.project) RIGHT(io.trino.sql.planner.plan.JoinNode.Type.RIGHT) NullLiteral(io.trino.sql.tree.NullLiteral) NullLiteral(io.trino.sql.tree.NullLiteral) BaseRuleTest(io.trino.sql.planner.iterative.rule.test.BaseRuleTest) Test(org.testng.annotations.Test)

Aggregations

ImmutableList (com.google.common.collect.ImmutableList)3 ImmutableMap (com.google.common.collect.ImmutableMap)3 FULL (io.trino.sql.planner.plan.JoinNode.Type.FULL)3 LEFT (io.trino.sql.planner.plan.JoinNode.Type.LEFT)3 RIGHT (io.trino.sql.planner.plan.JoinNode.Type.RIGHT)3 PlanNodeStatsEstimate (io.trino.cost.PlanNodeStatsEstimate)2 Symbol (io.trino.sql.planner.Symbol)2 JoinNode (io.trino.sql.planner.plan.JoinNode)2 TableScanNode (io.trino.sql.planner.plan.TableScanNode)2 Optional (java.util.Optional)2 Verify.verify (com.google.common.base.Verify.verify)1 ImmutableList.toImmutableList (com.google.common.collect.ImmutableList.toImmutableList)1 ImmutableListMultimap (com.google.common.collect.ImmutableListMultimap)1 ImmutableMap.toImmutableMap (com.google.common.collect.ImmutableMap.toImmutableMap)1 ImmutableSet (com.google.common.collect.ImmutableSet)1 Session (io.trino.Session)1 JOIN_DISTRIBUTION_TYPE (io.trino.SystemSessionProperties.JOIN_DISTRIBUTION_TYPE)1 JOIN_MAX_BROADCAST_TABLE_SIZE (io.trino.SystemSessionProperties.JOIN_MAX_BROADCAST_TABLE_SIZE)1 SystemSessionProperties.isAllowPushdownIntoConnectors (io.trino.SystemSessionProperties.isAllowPushdownIntoConnectors)1 CostComparator (io.trino.cost.CostComparator)1