StreamSerDe.java

/*
 * Copyright © 2014 Cask Data, Inc.
 *
 * 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 co.cask.cdap.hive.stream;

import co.cask.cdap.api.data.format.FormatSpecification;
import co.cask.cdap.api.data.schema.Schema;
import co.cask.cdap.api.data.schema.UnsupportedTypeException;
import co.cask.cdap.api.flow.flowlet.StreamEvent;
import co.cask.cdap.common.conf.Constants;
import co.cask.cdap.data2.transaction.stream.StreamConfig;
import co.cask.cdap.format.RecordFormats;
import co.cask.cdap.hive.context.ContextManager;
import co.cask.cdap.hive.serde.ObjectDeserializer;
import co.cask.cdap.internal.io.SchemaTypeAdapter;
import co.cask.cdap.proto.Id;
import co.cask.cdap.spi.stream.AbstractStreamEventRecordFormat;
import com.google.common.collect.Lists;
import com.google.gson.Gson;
import com.google.gson.GsonBuilder;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hive.serde2.SerDe;
import org.apache.hadoop.hive.serde2.SerDeException;
import org.apache.hadoop.hive.serde2.SerDeStats;
import org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector;
import org.apache.hadoop.io.ObjectWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.io.Writable;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import java.io.IOException;
import java.util.List;
import java.util.Properties;

/**
 * SerDe to deserialize Stream Events. It MUST implement the deprecated SerDe interface instead of extending the
 * abstract SerDe class, otherwise we get ClassNotFound exceptions on cdh4.x.
 */
@SuppressWarnings("deprecation")
public class StreamSerDe implements SerDe {
  private static final Logger LOG = LoggerFactory.getLogger(StreamSerDe.class);
  // timestamp and headers are guaranteed to be the first columns in a stream table.
  // the rest of the columns are for the stream body.
  private static final int BODY_OFFSET = 2;

  // A GSON object that knowns how to serialize Schema type.
  private static final Gson GSON = new GsonBuilder()
    .registerTypeAdapter(Schema.class, new SchemaTypeAdapter())
    .create();

  private ObjectInspector inspector;
  private AbstractStreamEventRecordFormat<?> streamFormat;
  private ObjectDeserializer deserializer;

  // initialize gets called multiple times by Hive. It may seem like a good idea to put additional settings into
  // the conf, but be very careful when doing so. If there are multiple hive tables involved in a query, initialize
  // for each table is called before input splits are fetched for any table. It is therefore not safe to put anything
  // the input format may need into conf in this method. Rather, use StorageHandler's method to place needed config
  // into the properties map there, which will get passed here and also copied into the job conf for the input
  // format to consume.
  @Override
  public void initialize(Configuration conf, Properties properties) throws SerDeException {
    // The columns property comes from the Hive metastore, which has it from the create table statement
    // It is then important that this schema be accurate and in the right order - the same order as
    // object inspectors will reflect them.

    String streamName = properties.getProperty(Constants.Explore.STREAM_NAME);
    String streamNamespace = properties.getProperty(Constants.Explore.STREAM_NAMESPACE);

    // no namespace SHOULD be an exception but... Hive calls initialize in several places, one of which is
    // when you try and drop a table.
    // When updating to CDAP 2.8, old tables will not have namespace as a serde property. So in order
    // to avoid a null pointer exception that prevents dropping a table, we handle the null namespace case here.
    if (streamNamespace == null) {
      // we also still need an ObjectInspector as Hive uses it to check what columns the table has.
      this.inspector = new ObjectDeserializer(properties, null).getInspector();
      return;
    }

    Id.Stream streamId = Id.Stream.from(streamNamespace, streamName);
    try (ContextManager.Context context = ContextManager.getContext(conf)) {
      Schema schema = null;
      // apparently the conf can be null in some versions of Hive?
      // Because it calls initialize just to get the object inspector
      if (context != null) {
        // Get the stream format from the stream config.
        FormatSpecification formatSpec = getFormatSpec(properties, streamId, context);
        this.streamFormat = (AbstractStreamEventRecordFormat) RecordFormats.createInitializedFormat(formatSpec);
        schema = formatSpec.getSchema();
      }
      this.deserializer = new ObjectDeserializer(properties, schema, BODY_OFFSET);
      this.inspector = deserializer.getInspector();
    } catch (UnsupportedTypeException e) {
      // this should have been validated up front when schema was set on the stream.
      // if we hit this something went wrong much earlier.
      LOG.error("Schema unsupported by format.", e);
      throw new SerDeException("Schema unsupported by format.", e);
    } catch (IOException e) {
      LOG.error("Could not get the config for stream {}.", streamName, e);
      throw new SerDeException("Could not get the config for stream " + streamName, e);
    } catch (Exception e) {
      LOG.error("Could not create the format for stream {}.", streamName, e);
      throw new SerDeException("Could not create the format for stream " + streamName, e);
    }
  }

  @Override
  public Class<? extends Writable> getSerializedClass() {
    return Text.class;
  }

  @Override
  public Writable serialize(Object o, ObjectInspector objectInspector) throws SerDeException {
    // should not be writing to streams through this
    throw new SerDeException("Stream serialization through Hive is not supported.");
  }

  @Override
  public SerDeStats getSerDeStats() {
    return new SerDeStats();
  }

  @Override
  public Object deserialize(Writable writable) throws SerDeException {
    // The writable should always contains a StreamEvent object provided by the StreamRecordReader
    ObjectWritable objectWritable = (ObjectWritable) writable;
    StreamEvent streamEvent = (StreamEvent) objectWritable.get();

    // timestamp and headers are always guaranteed to be first.
    List<Object> event = Lists.newArrayList();
    event.add(streamEvent.getTimestamp());
    event.add(streamEvent.getHeaders());

    try {
      // The format should always format the stream event into a record.
      event.addAll(deserializer.translateRecord(streamFormat.read(streamEvent)));
      return event;
    } catch (Throwable t) {
      LOG.info("Unable to format the stream body.", t);
      throw new SerDeException("Unable to format the stream body.", t);
    }
  }

  @Override
  public ObjectInspector getObjectInspector() throws SerDeException {
    return inspector;
  }

  /**
   * Gets the {@link FormatSpecification} for the given stream based on the SerDe properties.
   * For backward compatibility, if the format specification is not set in the SerDe properties, it will be
   * fetched from the {@link StreamConfig}.
   */
  private FormatSpecification getFormatSpec(Properties properties,
                                            Id.Stream streamId, ContextManager.Context context) throws IOException {
    String formatSpec = properties.getProperty(Constants.Explore.FORMAT_SPEC);
    if (formatSpec == null) {
      StreamConfig config = context.getStreamConfig(streamId);
      return config.getFormat();
    }
    return GSON.fromJson(formatSpec, FormatSpecification.class);
  }
}