path: root/exec/java-exec/src/main/java/org/apache/drill/exec/store/dfs/easy/EasyFormatPlugin.java

/*
 * Licensed to the Apache Software Foundation (ASF) under one
 * or more contributor license agreements.  See the NOTICE file
 * distributed with this work for additional information
 * regarding copyright ownership.  The ASF licenses this file
 * to you 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 org.apache.drill.exec.store.dfs.easy;

import java.io.IOException;
import java.util.LinkedList;
import java.util.List;
import java.util.Map;
import java.util.Set;

import org.apache.drill.shaded.guava.com.google.common.base.Functions;
import org.apache.drill.shaded.guava.com.google.common.collect.Maps;
import org.apache.drill.common.exceptions.ExecutionSetupException;
import org.apache.drill.common.expression.SchemaPath;
import org.apache.drill.common.logical.FormatPluginConfig;
import org.apache.drill.common.logical.StoragePluginConfig;
import org.apache.drill.exec.ops.FragmentContext;
import org.apache.drill.exec.ops.OperatorContext;
import org.apache.drill.exec.physical.base.AbstractGroupScan;
import org.apache.drill.exec.physical.base.AbstractWriter;
import org.apache.drill.exec.physical.base.PhysicalOperator;
import org.apache.drill.exec.physical.base.ScanStats;
import org.apache.drill.exec.physical.base.ScanStats.GroupScanProperty;
import org.apache.drill.exec.physical.impl.ScanBatch;
import org.apache.drill.exec.physical.impl.WriterRecordBatch;
import org.apache.drill.exec.physical.impl.StatisticsWriterRecordBatch;
import org.apache.drill.exec.planner.physical.PlannerSettings;
import org.apache.drill.exec.record.CloseableRecordBatch;
import org.apache.drill.exec.record.RecordBatch;
import org.apache.drill.exec.server.DrillbitContext;
import org.apache.drill.exec.store.ColumnExplorer;
import org.apache.drill.exec.store.RecordReader;
import org.apache.drill.exec.store.RecordWriter;
import org.apache.drill.exec.store.StatisticsRecordWriter;
import org.apache.drill.exec.store.StoragePluginOptimizerRule;
import org.apache.drill.exec.store.dfs.BasicFormatMatcher;
import org.apache.drill.exec.store.dfs.DrillFileSystem;
import org.apache.drill.exec.store.dfs.FileSelection;
import org.apache.drill.exec.store.dfs.FormatMatcher;
import org.apache.drill.exec.store.dfs.FormatPlugin;
import org.apache.drill.exec.store.schedule.CompleteFileWork;
import org.apache.hadoop.conf.Configuration;

import org.apache.drill.shaded.guava.com.google.common.collect.ImmutableSet;
import org.apache.drill.shaded.guava.com.google.common.collect.Lists;

public abstract class EasyFormatPlugin<T extends FormatPluginConfig> implements FormatPlugin {

  @SuppressWarnings("unused")
  private static final org.slf4j.Logger logger = org.slf4j.LoggerFactory.getLogger(EasyFormatPlugin.class);

  private final BasicFormatMatcher matcher;
  private final DrillbitContext context;
  private final boolean readable;
  private final boolean writable;
  private final boolean blockSplittable;
  private final Configuration fsConf;
  private final StoragePluginConfig storageConfig;
  protected final T formatConfig;
  private final String name;
  private final boolean compressible;

  protected EasyFormatPlugin(String name, DrillbitContext context, Configuration fsConf,
      StoragePluginConfig storageConfig, T formatConfig, boolean readable, boolean writable, boolean blockSplittable,
      boolean compressible, List<String> extensions, String defaultName){
    this.matcher = new BasicFormatMatcher(this, fsConf, extensions, compressible);
    this.readable = readable;
    this.writable = writable;
    this.context = context;
    this.blockSplittable = blockSplittable;
    this.compressible = compressible;
    this.fsConf = fsConf;
    this.storageConfig = storageConfig;
    this.formatConfig = formatConfig;
    this.name = name == null ? defaultName : name;
  }

  @Override
  public Configuration getFsConf() {
    return fsConf;
  }

  @Override
  public DrillbitContext getContext() {
    return context;
  }

  @Override
  public String getName() {
    return name;
  }

  public abstract boolean supportsPushDown();

  /**
   * Whether or not you can split the format based on blocks within file boundaries. If not, the simple format engine will
   * only split on file boundaries.
   *
   * @return True if splittable.
   */
  public boolean isBlockSplittable() {
    return blockSplittable;
  }

  /**
   * Indicates whether or not this format could also be in a compression
   * container (for example: csv.gz versus csv). If this format uses its own
   * internal compression scheme, such as Parquet does, then this should return
   * false.
   *
   * @return <code>true</code> if it is compressible
   */
  public boolean isCompressible() {
    return compressible;
  }

  public abstract RecordReader getRecordReader(FragmentContext context, DrillFileSystem dfs, FileWork fileWork,
      List<SchemaPath> columns, String userName) throws ExecutionSetupException;

  CloseableRecordBatch getReaderBatch(FragmentContext context, EasySubScan scan) throws ExecutionSetupException {
    final ColumnExplorer columnExplorer = new ColumnExplorer(context.getOptions(), scan.getColumns());

    if (!columnExplorer.isStarQuery()) {
      scan = new EasySubScan(scan.getUserName(), scan.getWorkUnits(), scan.getFormatPlugin(),
          columnExplorer.getTableColumns(), scan.getSelectionRoot());
      scan.setOperatorId(scan.getOperatorId());
    }

    OperatorContext oContext = context.newOperatorContext(scan);
    final DrillFileSystem dfs;
    try {
      dfs = oContext.newFileSystem(fsConf);
    } catch (IOException e) {
      throw new ExecutionSetupException(String.format("Failed to create FileSystem: %s", e.getMessage()), e);
    }

    List<RecordReader> readers = new LinkedList<>();
    List<Map<String, String>> implicitColumns = Lists.newArrayList();
    Map<String, String> mapWithMaxColumns = Maps.newLinkedHashMap();
    boolean supportsFileImplicitColumns = scan.getSelectionRoot() != null;
    for (FileWork work : scan.getWorkUnits()){
      RecordReader recordReader = getRecordReader(context, dfs, work, scan.getColumns(), scan.getUserName());
      readers.add(recordReader);
      List<String> partitionValues = ColumnExplorer.listPartitionValues(work.getPath(), scan.getSelectionRoot(), false);
      Map<String, String> implicitValues = columnExplorer.populateImplicitColumns(work.getPath(), partitionValues, supportsFileImplicitColumns);
      implicitColumns.add(implicitValues);
      if (implicitValues.size() > mapWithMaxColumns.size()) {
        mapWithMaxColumns = implicitValues;
      }
    }

    // all readers should have the same number of implicit columns, add missing ones with value null
    Map<String, String> diff = Maps.transformValues(mapWithMaxColumns, Functions.constant((String) null));
    for (Map<String, String> map : implicitColumns) {
      map.putAll(Maps.difference(map, diff).entriesOnlyOnRight());
    }

    return new ScanBatch(context, oContext, readers, implicitColumns);
  }

  public boolean isStatisticsRecordWriter(FragmentContext context, EasyWriter writer) {
    return false;
  }

  public abstract RecordWriter getRecordWriter(FragmentContext context, EasyWriter writer) throws IOException;

  public StatisticsRecordWriter getStatisticsRecordWriter(FragmentContext context, EasyWriter writer) throws IOException
  {
    return null;
  }

  public CloseableRecordBatch getWriterBatch(FragmentContext context, RecordBatch incoming, EasyWriter writer)
      throws ExecutionSetupException {
    try {
      if (isStatisticsRecordWriter(context, writer)) {
        return new StatisticsWriterRecordBatch(writer, incoming, context, getStatisticsRecordWriter(context, writer));
      } else {
        return new WriterRecordBatch(writer, incoming, context, getRecordWriter(context, writer));
      }
    } catch(IOException e) {
      throw new ExecutionSetupException(String.format("Failed to create the WriterRecordBatch. %s", e.getMessage()), e);
    }
  }

  protected ScanStats getScanStats(final PlannerSettings settings, final EasyGroupScan scan) {
    long data = 0;
    for (final CompleteFileWork work : scan.getWorkIterable()) {
      data += work.getTotalBytes();
    }

    final long estRowCount = data / 1024;
    return new ScanStats(GroupScanProperty.NO_EXACT_ROW_COUNT, estRowCount, 1, data);
  }

  @Override
  public AbstractWriter getWriter(PhysicalOperator child, String location, List<String> partitionColumns) throws IOException {
    return new EasyWriter(child, location, partitionColumns, this);
  }

  @Override
  public AbstractGroupScan getGroupScan(String userName, FileSelection selection, List<SchemaPath> columns)
      throws IOException {
    return new EasyGroupScan(userName, selection, this, columns, selection.selectionRoot);
  }

  @Override
  public T getConfig() {
    return formatConfig;
  }

  @Override
  public StoragePluginConfig getStorageConfig() {
    return storageConfig;
  }

  @Override
  public boolean supportsRead() {
    return readable;
  }

  @Override
  public boolean supportsWrite() {
    return writable;
  }

  @Override
  public boolean supportsAutoPartitioning() {
    return false;
  }

  @Override
  public FormatMatcher getMatcher() {
    return matcher;
  }

  @Override
  public Set<StoragePluginOptimizerRule> getOptimizerRules() {
    return ImmutableSet.of();
  }

  public abstract int getReaderOperatorType();
  public abstract int getWriterOperatorType();

}