All Downloads are FREE. Search and download functionalities are using the official Maven repository.

io.trino.plugin.hive.HivePageSourceProvider Maven / Gradle / Ivy

Go to download

This is a Databricks build of Trino's Hive plugin which includes support for HTTP based transport for it's Hive metastore thrift interface.

The newest version!
/*
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
package io.trino.plugin.hive;

import com.google.common.annotations.VisibleForTesting;
import com.google.common.collect.BiMap;
import com.google.common.collect.ImmutableBiMap;
import com.google.common.collect.ImmutableList;
import com.google.common.collect.ImmutableSet;
import io.trino.hdfs.HdfsContext;
import io.trino.hdfs.HdfsEnvironment;
import io.trino.plugin.hive.HivePageSource.BucketValidator;
import io.trino.plugin.hive.HiveRecordCursorProvider.ReaderRecordCursorWithProjections;
import io.trino.plugin.hive.HiveSplit.BucketConversion;
import io.trino.plugin.hive.HiveSplit.BucketValidation;
import io.trino.plugin.hive.acid.AcidTransaction;
import io.trino.plugin.hive.type.TypeInfo;
import io.trino.plugin.hive.util.HiveBucketing.BucketingVersion;
import io.trino.spi.connector.ColumnHandle;
import io.trino.spi.connector.ConnectorPageSource;
import io.trino.spi.connector.ConnectorPageSourceProvider;
import io.trino.spi.connector.ConnectorSession;
import io.trino.spi.connector.ConnectorSplit;
import io.trino.spi.connector.ConnectorTableHandle;
import io.trino.spi.connector.ConnectorTransactionHandle;
import io.trino.spi.connector.DynamicFilter;
import io.trino.spi.connector.EmptyPageSource;
import io.trino.spi.connector.RecordCursor;
import io.trino.spi.connector.RecordPageSource;
import io.trino.spi.predicate.Domain;
import io.trino.spi.predicate.NullableValue;
import io.trino.spi.predicate.TupleDomain;
import io.trino.spi.type.Type;
import io.trino.spi.type.TypeManager;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;

import javax.inject.Inject;

import java.util.ArrayList;
import java.util.HashMap;
import java.util.HashSet;
import java.util.List;
import java.util.Map;
import java.util.Objects;
import java.util.Optional;
import java.util.OptionalInt;
import java.util.Properties;
import java.util.Set;

import static com.google.common.base.Preconditions.checkArgument;
import static com.google.common.base.Preconditions.checkState;
import static com.google.common.collect.ImmutableList.toImmutableList;
import static com.google.common.collect.ImmutableMap.toImmutableMap;
import static com.google.common.collect.Maps.uniqueIndex;
import static io.trino.plugin.hive.AbstractHiveAcidWriters.ORIGINAL_FILE_PATH_MATCHER;
import static io.trino.plugin.hive.HiveColumnHandle.ColumnType.PARTITION_KEY;
import static io.trino.plugin.hive.HiveColumnHandle.ColumnType.REGULAR;
import static io.trino.plugin.hive.HiveColumnHandle.ColumnType.SYNTHESIZED;
import static io.trino.plugin.hive.HiveColumnHandle.isRowIdColumnHandle;
import static io.trino.plugin.hive.HivePageSourceProvider.ColumnMapping.toColumnHandles;
import static io.trino.plugin.hive.HivePageSourceProvider.ColumnMappingKind.PREFILLED;
import static io.trino.plugin.hive.util.HiveBucketing.HiveBucketFilter;
import static io.trino.plugin.hive.util.HiveBucketing.getHiveBucketFilter;
import static io.trino.plugin.hive.util.HiveUtil.getPrefilledColumnValue;
import static java.util.Objects.requireNonNull;
import static java.util.function.Function.identity;
import static java.util.stream.Collectors.toList;

public class HivePageSourceProvider
        implements ConnectorPageSourceProvider
{
    private final TypeManager typeManager;
    private final HdfsEnvironment hdfsEnvironment;
    private final int domainCompactionThreshold;
    private final Set pageSourceFactories;
    private final Set cursorProviders;

    @Inject
    public HivePageSourceProvider(
            TypeManager typeManager,
            HdfsEnvironment hdfsEnvironment,
            HiveConfig hiveConfig,
            Set pageSourceFactories,
            Set cursorProviders,
            GenericHiveRecordCursorProvider genericCursorProvider)
    {
        this.typeManager = requireNonNull(typeManager, "typeManager is null");
        this.hdfsEnvironment = requireNonNull(hdfsEnvironment, "hdfsEnvironment is null");
        this.domainCompactionThreshold = hiveConfig.getDomainCompactionThreshold();
        this.pageSourceFactories = ImmutableSet.copyOf(requireNonNull(pageSourceFactories, "pageSourceFactories is null"));
        this.cursorProviders = ImmutableSet.builder()
                .addAll(requireNonNull(cursorProviders, "cursorProviders is null"))
                .add(genericCursorProvider) // generic should be last, as a fallback option
                .build();
    }

    @Override
    public ConnectorPageSource createPageSource(
            ConnectorTransactionHandle transaction,
            ConnectorSession session,
            ConnectorSplit split,
            ConnectorTableHandle tableHandle,
            List columns,
            DynamicFilter dynamicFilter)
    {
        HiveTableHandle hiveTable = (HiveTableHandle) tableHandle;
        HiveSplit hiveSplit = (HiveSplit) split;

        if (shouldSkipBucket(hiveTable, hiveSplit, dynamicFilter)) {
            return new EmptyPageSource();
        }

        List hiveColumns = columns.stream()
                .map(HiveColumnHandle.class::cast)
                .collect(toList());

        Path path = new Path(hiveSplit.getPath());
        boolean originalFile = ORIGINAL_FILE_PATH_MATCHER.matcher(path.toString()).matches();

        List columnMappings = ColumnMapping.buildColumnMappings(
                hiveSplit.getPartitionName(),
                hiveSplit.getPartitionKeys(),
                hiveColumns,
                hiveSplit.getBucketConversion().map(BucketConversion::getBucketColumnHandles).orElse(ImmutableList.of()),
                hiveSplit.getTableToPartitionMapping(),
                path,
                hiveSplit.getTableBucketNumber(),
                hiveSplit.getEstimatedFileSize(),
                hiveSplit.getFileModifiedTime());

        // Perform dynamic partition pruning in case coordinator didn't prune split.
        // This can happen when dynamic filters are collected after partition splits were listed.
        if (shouldSkipSplit(columnMappings, dynamicFilter)) {
            return new EmptyPageSource();
        }

        Configuration configuration = hdfsEnvironment.getConfiguration(new HdfsContext(session), path);

        TupleDomain simplifiedDynamicFilter = dynamicFilter
                .getCurrentPredicate()
                .transformKeys(HiveColumnHandle.class::cast).simplify(domainCompactionThreshold);
        Optional pageSource = createHivePageSource(
                pageSourceFactories,
                cursorProviders,
                configuration,
                session,
                path,
                hiveSplit.getTableBucketNumber(),
                hiveSplit.getStart(),
                hiveSplit.getLength(),
                hiveSplit.getEstimatedFileSize(),
                hiveSplit.getSchema(),
                hiveTable.getCompactEffectivePredicate().intersect(simplifiedDynamicFilter),
                hiveColumns,
                typeManager,
                hiveSplit.getBucketConversion(),
                hiveSplit.getBucketValidation(),
                hiveSplit.isS3SelectPushdownEnabled(),
                hiveSplit.getAcidInfo(),
                originalFile,
                hiveTable.getTransaction(),
                columnMappings);

        if (pageSource.isPresent()) {
            return pageSource.get();
        }
        throw new RuntimeException("Could not find a file reader for split " + hiveSplit);
    }

    public static Optional createHivePageSource(
            Set pageSourceFactories,
            Set cursorProviders,
            Configuration configuration,
            ConnectorSession session,
            Path path,
            OptionalInt tableBucketNumber,
            long start,
            long length,
            long estimatedFileSize,
            Properties schema,
            TupleDomain effectivePredicate,
            List columns,
            TypeManager typeManager,
            Optional bucketConversion,
            Optional bucketValidation,
            boolean s3SelectPushdownEnabled,
            Optional acidInfo,
            boolean originalFile,
            AcidTransaction transaction,
            List columnMappings)
    {
        if (effectivePredicate.isNone()) {
            return Optional.of(new EmptyPageSource());
        }

        List regularAndInterimColumnMappings = ColumnMapping.extractRegularAndInterimColumnMappings(columnMappings);

        Optional bucketAdaptation = createBucketAdaptation(bucketConversion, tableBucketNumber, regularAndInterimColumnMappings);
        Optional bucketValidator = createBucketValidator(path, bucketValidation, tableBucketNumber, regularAndInterimColumnMappings);

        for (HivePageSourceFactory pageSourceFactory : pageSourceFactories) {
            List desiredColumns = toColumnHandles(regularAndInterimColumnMappings, true, typeManager);

            Optional readerWithProjections = pageSourceFactory.createPageSource(
                    configuration,
                    session,
                    path,
                    start,
                    length,
                    estimatedFileSize,
                    schema,
                    desiredColumns,
                    effectivePredicate,
                    acidInfo,
                    tableBucketNumber,
                    originalFile,
                    transaction);

            if (readerWithProjections.isPresent()) {
                ConnectorPageSource pageSource = readerWithProjections.get().get();

                Optional readerProjections = readerWithProjections.get().getReaderColumns();
                Optional adapter = Optional.empty();
                if (readerProjections.isPresent()) {
                    adapter = Optional.of(hiveProjectionsAdapter(desiredColumns, readerProjections.get()));
                }

                return Optional.of(new HivePageSource(
                        columnMappings,
                        bucketAdaptation,
                        bucketValidator,
                        adapter,
                        typeManager,
                        pageSource));
            }
        }

        for (HiveRecordCursorProvider provider : cursorProviders) {
            // GenericHiveRecordCursor will automatically do the coercion without HiveCoercionRecordCursor
            boolean doCoercion = !(provider instanceof GenericHiveRecordCursorProvider);

            List desiredColumns = toColumnHandles(regularAndInterimColumnMappings, doCoercion, typeManager);
            Optional readerWithProjections = provider.createRecordCursor(
                    configuration,
                    session,
                    path,
                    start,
                    length,
                    estimatedFileSize,
                    schema,
                    desiredColumns,
                    effectivePredicate,
                    typeManager,
                    s3SelectPushdownEnabled);

            if (readerWithProjections.isPresent()) {
                RecordCursor delegate = readerWithProjections.get().getRecordCursor();
                Optional projections = readerWithProjections.get().getProjectedReaderColumns();

                if (projections.isPresent()) {
                    ReaderProjectionsAdapter projectionsAdapter = hiveProjectionsAdapter(desiredColumns, projections.get());
                    delegate = new HiveReaderProjectionsAdaptingRecordCursor(delegate, projectionsAdapter);
                }

                checkArgument(acidInfo.isEmpty(), "Acid is not supported");

                if (bucketAdaptation.isPresent()) {
                    delegate = new HiveBucketAdapterRecordCursor(
                            bucketAdaptation.get().getBucketColumnIndices(),
                            bucketAdaptation.get().getBucketColumnHiveTypes(),
                            bucketAdaptation.get().getBucketingVersion(),
                            bucketAdaptation.get().getTableBucketCount(),
                            bucketAdaptation.get().getPartitionBucketCount(),
                            bucketAdaptation.get().getBucketToKeep(),
                            typeManager,
                            delegate);
                }

                // Need to wrap RcText and RcBinary into a wrapper, which will do the coercion for mismatch columns
                if (doCoercion) {
                    delegate = new HiveCoercionRecordCursor(regularAndInterimColumnMappings, typeManager, delegate);
                }

                // bucket adaptation already validates that data is in the right bucket
                if (bucketAdaptation.isEmpty() && bucketValidator.isPresent()) {
                    delegate = bucketValidator.get().wrapRecordCursor(delegate, typeManager);
                }

                HiveRecordCursor hiveRecordCursor = new HiveRecordCursor(columnMappings, delegate);
                List columnTypes = columns.stream()
                        .map(HiveColumnHandle::getType)
                        .collect(toList());

                return Optional.of(new RecordPageSource(columnTypes, hiveRecordCursor));
            }
        }

        return Optional.empty();
    }

    private static boolean shouldSkipBucket(HiveTableHandle hiveTable, HiveSplit hiveSplit, DynamicFilter dynamicFilter)
    {
        if (hiveSplit.getTableBucketNumber().isEmpty()) {
            return false;
        }
        Optional hiveBucketFilter = getHiveBucketFilter(hiveTable, dynamicFilter.getCurrentPredicate());
        return hiveBucketFilter.map(filter -> !filter.getBucketsToKeep().contains(hiveSplit.getTableBucketNumber().getAsInt())).orElse(false);
    }

    private static boolean shouldSkipSplit(List columnMappings, DynamicFilter dynamicFilter)
    {
        TupleDomain predicate = dynamicFilter.getCurrentPredicate();
        if (predicate.isNone()) {
            return true;
        }
        Map domains = predicate.getDomains().get();
        for (ColumnMapping columnMapping : columnMappings) {
            if (columnMapping.getKind() != PREFILLED) {
                continue;
            }
            Object value = columnMapping.getPrefilledValue().getValue();
            Domain allowedDomain = domains.get(columnMapping.getHiveColumnHandle());
            if (allowedDomain != null && !allowedDomain.includesNullableValue(value)) {
                return true;
            }
        }
        return false;
    }

    private static ReaderProjectionsAdapter hiveProjectionsAdapter(List expectedColumns, ReaderColumns readColumns)
    {
        return new ReaderProjectionsAdapter(
                expectedColumns.stream().map(ColumnHandle.class::cast).collect(toImmutableList()),
                readColumns,
                column -> ((HiveColumnHandle) column).getType(),
                HivePageSourceProvider::getProjection);
    }

    @VisibleForTesting
    static List getProjection(ColumnHandle expected, ColumnHandle read)
    {
        HiveColumnHandle expectedColumn = (HiveColumnHandle) expected;
        HiveColumnHandle readColumn = (HiveColumnHandle) read;

        checkArgument(expectedColumn.getBaseColumn().equals(readColumn.getBaseColumn()), "reader column is not valid for expected column");

        List expectedDereferences = expectedColumn.getHiveColumnProjectionInfo()
                .map(HiveColumnProjectionInfo::getDereferenceIndices)
                .orElse(ImmutableList.of());

        List readerDereferences = readColumn.getHiveColumnProjectionInfo()
                .map(HiveColumnProjectionInfo::getDereferenceIndices)
                .orElse(ImmutableList.of());

        checkArgument(readerDereferences.size() <= expectedDereferences.size(), "Field returned by the reader should include expected field");
        checkArgument(expectedDereferences.subList(0, readerDereferences.size()).equals(readerDereferences), "Field returned by the reader should be a prefix of expected field");

        return expectedDereferences.subList(readerDereferences.size(), expectedDereferences.size());
    }

    public static class ColumnMapping
    {
        private final ColumnMappingKind kind;
        private final HiveColumnHandle hiveColumnHandle;
        private final Optional prefilledValue;
        /**
         * ordinal of this column in the underlying page source or record cursor
         */
        private final OptionalInt index;
        private final Optional baseTypeCoercionFrom;

        public static ColumnMapping regular(HiveColumnHandle hiveColumnHandle, int index, Optional baseTypeCoercionFrom)
        {
            checkArgument(hiveColumnHandle.getColumnType() == REGULAR);
            return new ColumnMapping(ColumnMappingKind.REGULAR, hiveColumnHandle, Optional.empty(), OptionalInt.of(index), baseTypeCoercionFrom);
        }

        public static ColumnMapping synthesized(HiveColumnHandle hiveColumnHandle, int index, Optional baseTypeCoercionFrom)
        {
            checkArgument(hiveColumnHandle.getColumnType() == SYNTHESIZED);
            return new ColumnMapping(ColumnMappingKind.SYNTHESIZED, hiveColumnHandle, Optional.empty(), OptionalInt.of(index), baseTypeCoercionFrom);
        }

        public static ColumnMapping prefilled(HiveColumnHandle hiveColumnHandle, NullableValue prefilledValue, Optional baseTypeCoercionFrom)
        {
            checkArgument(hiveColumnHandle.getColumnType() == PARTITION_KEY || hiveColumnHandle.getColumnType() == SYNTHESIZED);
            checkArgument(hiveColumnHandle.isBaseColumn(), "prefilled values not supported for projected columns");
            return new ColumnMapping(PREFILLED, hiveColumnHandle, Optional.of(prefilledValue), OptionalInt.empty(), baseTypeCoercionFrom);
        }

        public static ColumnMapping interim(HiveColumnHandle hiveColumnHandle, int index, Optional baseTypeCoercionFrom)
        {
            checkArgument(hiveColumnHandle.getColumnType() == REGULAR);
            return new ColumnMapping(ColumnMappingKind.INTERIM, hiveColumnHandle, Optional.empty(), OptionalInt.of(index), baseTypeCoercionFrom);
        }

        public static ColumnMapping empty(HiveColumnHandle hiveColumnHandle)
        {
            checkArgument(hiveColumnHandle.getColumnType() == REGULAR);
            return new ColumnMapping(ColumnMappingKind.EMPTY, hiveColumnHandle, Optional.empty(), OptionalInt.empty(), Optional.empty());
        }

        private ColumnMapping(
                ColumnMappingKind kind,
                HiveColumnHandle hiveColumnHandle,
                Optional prefilledValue,
                OptionalInt index,
                Optional baseTypeCoercionFrom)
        {
            this.kind = requireNonNull(kind, "kind is null");
            this.hiveColumnHandle = requireNonNull(hiveColumnHandle, "hiveColumnHandle is null");
            this.prefilledValue = requireNonNull(prefilledValue, "prefilledValue is null");
            this.index = requireNonNull(index, "index is null");
            this.baseTypeCoercionFrom = requireNonNull(baseTypeCoercionFrom, "baseTypeCoercionFrom is null");
        }

        public ColumnMappingKind getKind()
        {
            return kind;
        }

        public NullableValue getPrefilledValue()
        {
            checkState(kind == PREFILLED);
            return prefilledValue.get();
        }

        public HiveColumnHandle getHiveColumnHandle()
        {
            return hiveColumnHandle;
        }

        public int getIndex()
        {
            checkState(kind == ColumnMappingKind.REGULAR || kind == ColumnMappingKind.INTERIM || isRowIdColumnHandle(hiveColumnHandle));
            return index.getAsInt();
        }

        public Optional getBaseTypeCoercionFrom()
        {
            return baseTypeCoercionFrom;
        }

        public static List buildColumnMappings(
                String partitionName,
                List partitionKeys,
                List columns,
                List requiredInterimColumns,
                TableToPartitionMapping tableToPartitionMapping,
                Path path,
                OptionalInt bucketNumber,
                long estimatedFileSize,
                long fileModifiedTime)
        {
            Map partitionKeysByName = uniqueIndex(partitionKeys, HivePartitionKey::getName);

            // Maintain state about hive columns added to the mapping as we iterate (for validation)
            Set baseColumnHiveIndices = new HashSet<>();
            Map>> projectionsForColumn = new HashMap<>();

            ImmutableList.Builder columnMappings = ImmutableList.builder();
            int regularIndex = 0;

            for (HiveColumnHandle column : columns) {
                Optional baseTypeCoercionFrom = tableToPartitionMapping.getCoercion(column.getBaseHiveColumnIndex());
                if (column.getColumnType() == REGULAR) {
                    if (column.isBaseColumn()) {
                        baseColumnHiveIndices.add(column.getBaseHiveColumnIndex());
                    }

                    checkArgument(
                            projectionsForColumn.computeIfAbsent(column.getBaseHiveColumnIndex(), columnIndex -> new HashSet<>()).add(column.getHiveColumnProjectionInfo()),
                            "duplicate column in columns list");

                    // Add regular mapping if projection is valid for partition schema, otherwise add an empty mapping
                    if (baseTypeCoercionFrom.isEmpty()
                            || projectionValidForType(baseTypeCoercionFrom.get(), column.getHiveColumnProjectionInfo())) {
                        columnMappings.add(regular(column, regularIndex, baseTypeCoercionFrom));
                        regularIndex++;
                    }
                    else {
                        columnMappings.add(empty(column));
                    }
                }
                else if (isRowIdColumnHandle(column)) {
                    baseColumnHiveIndices.add(column.getBaseHiveColumnIndex());
                    checkArgument(
                            projectionsForColumn.computeIfAbsent(column.getBaseHiveColumnIndex(), index -> new HashSet<>()).add(column.getHiveColumnProjectionInfo()),
                            "duplicate column in columns list");

                    if (baseTypeCoercionFrom.isEmpty()
                            || projectionValidForType(baseTypeCoercionFrom.get(), column.getHiveColumnProjectionInfo())) {
                        columnMappings.add(synthesized(column, regularIndex, baseTypeCoercionFrom));
                    }
                    else {
                        throw new RuntimeException("baseTypeCoercisionFrom was empty for the rowId column");
                    }
                    regularIndex++;
                }
                else {
                    columnMappings.add(prefilled(
                            column,
                            getPrefilledColumnValue(column, partitionKeysByName.get(column.getName()), path, bucketNumber, estimatedFileSize, fileModifiedTime, partitionName),
                            baseTypeCoercionFrom));
                }
            }

            for (HiveColumnHandle column : requiredInterimColumns) {
                checkArgument(column.getColumnType() == REGULAR);
                checkArgument(column.isBaseColumn(), "bucketed columns should be base columns");
                if (baseColumnHiveIndices.contains(column.getBaseHiveColumnIndex())) {
                    continue; // This column exists in columns. Do not add it again.
                }

                if (projectionsForColumn.containsKey(column.getBaseHiveColumnIndex())) {
                    columnMappings.add(interim(column, regularIndex, tableToPartitionMapping.getCoercion(column.getBaseHiveColumnIndex())));
                }
                else {
                    // If coercion does not affect bucket number calculation, coercion doesn't need to be applied here.
                    // Otherwise, read of this partition should not be allowed.
                    // (Alternatively, the partition could be read as an unbucketed partition. This is not implemented.)
                    columnMappings.add(interim(column, regularIndex, Optional.empty()));
                }
                regularIndex++;
            }
            return columnMappings.build();
        }

        private static boolean projectionValidForType(HiveType baseType, Optional projection)
        {
            List dereferences = projection.map(HiveColumnProjectionInfo::getDereferenceIndices).orElse(ImmutableList.of());
            Optional targetType = baseType.getHiveTypeForDereferences(dereferences);
            return targetType.isPresent();
        }

        public static List extractRegularAndInterimColumnMappings(List columnMappings)
        {
            return columnMappings.stream()
                    .filter(columnMapping -> columnMapping.getKind() == ColumnMappingKind.REGULAR || columnMapping.getKind() == ColumnMappingKind.INTERIM)
                    .collect(toImmutableList());
        }

        public static List toColumnHandles(List regularColumnMappings, boolean doCoercion, TypeManager typeManager)
        {
            return regularColumnMappings.stream()
                    .map(columnMapping -> {
                        HiveColumnHandle columnHandle = columnMapping.getHiveColumnHandle();
                        if (!doCoercion || columnMapping.getBaseTypeCoercionFrom().isEmpty()) {
                            return columnHandle;
                        }
                        HiveType fromHiveTypeBase = columnMapping.getBaseTypeCoercionFrom().get();

                        Optional newColumnProjectionInfo = columnHandle.getHiveColumnProjectionInfo().map(projectedColumn -> {
                            HiveType fromHiveType = fromHiveTypeBase.getHiveTypeForDereferences(projectedColumn.getDereferenceIndices()).get();
                            return new HiveColumnProjectionInfo(
                                    projectedColumn.getDereferenceIndices(),
                                    projectedColumn.getDereferenceNames(),
                                    fromHiveType,
                                    fromHiveType.getType(typeManager));
                        });

                        return new HiveColumnHandle(
                                columnHandle.getBaseColumnName(),
                                columnHandle.getBaseHiveColumnIndex(),
                                fromHiveTypeBase,
                                fromHiveTypeBase.getType(typeManager),
                                newColumnProjectionInfo,
                                columnHandle.getColumnType(),
                                columnHandle.getComment());
                    })
                    .collect(toList());
        }
    }

    public enum ColumnMappingKind
    {
        REGULAR,
        PREFILLED,
        INTERIM,
        SYNTHESIZED,
        EMPTY
    }

    private static Optional createBucketAdaptation(Optional bucketConversion, OptionalInt bucketNumber, List columnMappings)
    {
        return bucketConversion.map(conversion -> {
            List baseColumnMapping = columnMappings.stream()
                    .filter(mapping -> mapping.getHiveColumnHandle().isBaseColumn())
                    .collect(toList());
            Map baseHiveColumnToBlockIndex = uniqueIndex(baseColumnMapping, mapping -> mapping.getHiveColumnHandle().getBaseHiveColumnIndex());

            int[] bucketColumnIndices = conversion.getBucketColumnHandles().stream()
                    .mapToInt(columnHandle -> baseHiveColumnToBlockIndex.get(columnHandle.getBaseHiveColumnIndex()).getIndex())
                    .toArray();
            List bucketColumnHiveTypes = conversion.getBucketColumnHandles().stream()
                    .map(columnHandle -> baseHiveColumnToBlockIndex.get(columnHandle.getBaseHiveColumnIndex()).getHiveColumnHandle().getHiveType())
                    .collect(toImmutableList());
            return new BucketAdaptation(
                    bucketColumnIndices,
                    bucketColumnHiveTypes,
                    conversion.getBucketingVersion(),
                    conversion.getTableBucketCount(),
                    conversion.getPartitionBucketCount(),
                    bucketNumber.getAsInt());
        });
    }

    public static class BucketAdaptation
    {
        private final int[] bucketColumnIndices;
        private final List bucketColumnHiveTypes;
        private final BucketingVersion bucketingVersion;
        private final int tableBucketCount;
        private final int partitionBucketCount;
        private final int bucketToKeep;

        public BucketAdaptation(
                int[] bucketColumnIndices,
                List bucketColumnHiveTypes,
                BucketingVersion bucketingVersion,
                int tableBucketCount,
                int partitionBucketCount,
                int bucketToKeep)
        {
            this.bucketColumnIndices = bucketColumnIndices;
            this.bucketColumnHiveTypes = bucketColumnHiveTypes;
            this.bucketingVersion = bucketingVersion;
            this.tableBucketCount = tableBucketCount;
            this.partitionBucketCount = partitionBucketCount;
            this.bucketToKeep = bucketToKeep;
        }

        public int[] getBucketColumnIndices()
        {
            return bucketColumnIndices;
        }

        public List getBucketColumnHiveTypes()
        {
            return bucketColumnHiveTypes;
        }

        public BucketingVersion getBucketingVersion()
        {
            return bucketingVersion;
        }

        public int getTableBucketCount()
        {
            return tableBucketCount;
        }

        public int getPartitionBucketCount()
        {
            return partitionBucketCount;
        }

        public int getBucketToKeep()
        {
            return bucketToKeep;
        }
    }

    private static Optional createBucketValidator(Path path, Optional bucketValidation, OptionalInt bucketNumber, List columnMappings)
    {
        return bucketValidation.flatMap(validation -> {
            Map baseHiveColumnToBlockIndex = columnMappings.stream()
                    .filter(mapping -> mapping.getHiveColumnHandle().isBaseColumn())
                    .collect(toImmutableMap(mapping -> mapping.getHiveColumnHandle().getBaseHiveColumnIndex(), identity()));

            int[] bucketColumnIndices = new int[validation.getBucketColumns().size()];

            List bucketColumnTypes = new ArrayList<>();
            for (int i = 0; i < validation.getBucketColumns().size(); i++) {
                HiveColumnHandle column = validation.getBucketColumns().get(i);
                ColumnMapping mapping = baseHiveColumnToBlockIndex.get(column.getBaseHiveColumnIndex());
                if (mapping == null) {
                    // The bucket column is not read by the query, and thus invalid bucketing cannot
                    // affect the results. Filtering on the hidden $bucket column still correctly
                    // partitions the table by bucket, even if the bucket has the wrong data.
                    return Optional.empty();
                }
                bucketColumnIndices[i] = mapping.getIndex();
                bucketColumnTypes.add(mapping.getHiveColumnHandle().getHiveType().getTypeInfo());
            }

            return Optional.of(new BucketValidator(
                    path,
                    bucketColumnIndices,
                    bucketColumnTypes,
                    validation.getBucketingVersion(),
                    validation.getBucketCount(),
                    bucketNumber.orElseThrow()));
        });
    }

    /**
     * Creates a mapping between the input {@param columns} and base columns if required.
     */
    public static Optional projectBaseColumns(List columns)
    {
        requireNonNull(columns, "columns is null");

        // No projection is required if all columns are base columns
        if (columns.stream().allMatch(HiveColumnHandle::isBaseColumn)) {
            return Optional.empty();
        }

        ImmutableList.Builder projectedColumns = ImmutableList.builder();
        ImmutableList.Builder outputColumnMapping = ImmutableList.builder();
        Map mappedHiveColumnIndices = new HashMap<>();
        int projectedColumnCount = 0;

        for (HiveColumnHandle column : columns) {
            int hiveColumnIndex = column.getBaseHiveColumnIndex();
            Integer mapped = mappedHiveColumnIndices.get(hiveColumnIndex);

            if (mapped == null) {
                projectedColumns.add(column.getBaseColumn());
                mappedHiveColumnIndices.put(hiveColumnIndex, projectedColumnCount);
                outputColumnMapping.add(projectedColumnCount);
                projectedColumnCount++;
            }
            else {
                outputColumnMapping.add(mapped);
            }
        }

        return Optional.of(new ReaderColumns(projectedColumns.build(), outputColumnMapping.build()));
    }

    /**
     * Creates a set of sufficient columns for the input projected columns and prepares a mapping between the two. For example,
     * if input {@param columns} include columns "a.b" and "a.b.c", then they will be projected from a single column "a.b".
     */
    public static Optional projectSufficientColumns(List columns)
    {
        requireNonNull(columns, "columns is null");

        if (columns.stream().allMatch(HiveColumnHandle::isBaseColumn)) {
            return Optional.empty();
        }

        ImmutableBiMap.Builder dereferenceChainsBuilder = ImmutableBiMap.builder();

        for (HiveColumnHandle column : columns) {
            List indices = column.getHiveColumnProjectionInfo()
                    .map(HiveColumnProjectionInfo::getDereferenceIndices)
                    .orElse(ImmutableList.of());

            DereferenceChain dereferenceChain = new DereferenceChain(column.getBaseColumnName(), indices);
            dereferenceChainsBuilder.put(dereferenceChain, column);
        }

        BiMap dereferenceChains = dereferenceChainsBuilder.build();

        List sufficientColumns = new ArrayList<>();
        ImmutableList.Builder outputColumnMapping = ImmutableList.builder();

        Map pickedColumns = new HashMap<>();

        // Pick a covering column for every column
        for (HiveColumnHandle columnHandle : columns) {
            DereferenceChain column = dereferenceChains.inverse().get(columnHandle);
            List orderedPrefixes = column.getOrderedPrefixes();
            DereferenceChain chosenColumn = null;

            // Shortest existing prefix is chosen as the input.
            for (DereferenceChain prefix : orderedPrefixes) {
                if (dereferenceChains.containsKey(prefix)) {
                    chosenColumn = prefix;
                    break;
                }
            }

            checkState(chosenColumn != null, "chosenColumn is null");
            int inputBlockIndex;

            if (pickedColumns.containsKey(chosenColumn)) {
                // Use already picked column
                inputBlockIndex = pickedColumns.get(chosenColumn);
            }
            else {
                // Add a new column for the reader
                sufficientColumns.add(dereferenceChains.get(chosenColumn));
                pickedColumns.put(chosenColumn, sufficientColumns.size() - 1);
                inputBlockIndex = sufficientColumns.size() - 1;
            }

            outputColumnMapping.add(inputBlockIndex);
        }

        return Optional.of(new ReaderColumns(sufficientColumns, outputColumnMapping.build()));
    }

    private static class DereferenceChain
    {
        private final String name;
        private final List indices;

        public DereferenceChain(String name, List indices)
        {
            this.name = requireNonNull(name, "name is null");
            this.indices = ImmutableList.copyOf(requireNonNull(indices, "indices is null"));
        }

        @Override
        public boolean equals(Object o)
        {
            if (this == o) {
                return true;
            }
            if (o == null || getClass() != o.getClass()) {
                return false;
            }

            DereferenceChain that = (DereferenceChain) o;
            return Objects.equals(name, that.name) &&
                    Objects.equals(indices, that.indices);
        }

        @Override
        public int hashCode()
        {
            return Objects.hash(name, indices);
        }

        /**
         * Get Prefixes of this Dereference chain in increasing order of lengths
         */
        public List getOrderedPrefixes()
        {
            ImmutableList.Builder prefixes = ImmutableList.builder();

            for (int prefixLen = 0; prefixLen <= indices.size(); prefixLen++) {
                prefixes.add(new DereferenceChain(name, indices.subList(0, prefixLen)));
            }

            return prefixes.build();
        }
    }
}




© 2015 - 2024 Weber Informatics LLC | Privacy Policy