drill-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From paul-rogers <...@git.apache.org>
Subject [GitHub] drill pull request #831: DRILL-5432: Added pcap-format support
Date Fri, 26 May 2017 01:44:17 GMT
Github user paul-rogers commented on a diff in the pull request:

    https://github.com/apache/drill/pull/831#discussion_r118617482
  
    --- Diff: exec/java-exec/src/main/java/org/apache/drill/exec/store/pcap/PcapRecordReader.java
---
    @@ -0,0 +1,295 @@
    +/*
    + * 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
    + * <p>
    + * http://www.apache.org/licenses/LICENSE-2.0
    + * <p>
    + * 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.pcap;
    +
    +import com.google.common.collect.ImmutableList;
    +import com.google.common.collect.ImmutableMap;
    +import org.apache.drill.common.exceptions.ExecutionSetupException;
    +import org.apache.drill.common.expression.SchemaPath;
    +import org.apache.drill.common.types.TypeProtos;
    +import org.apache.drill.common.types.TypeProtos.MajorType;
    +import org.apache.drill.common.types.TypeProtos.MinorType;
    +import org.apache.drill.common.types.Types;
    +import org.apache.drill.exec.exception.SchemaChangeException;
    +import org.apache.drill.exec.expr.TypeHelper;
    +import org.apache.drill.exec.ops.OperatorContext;
    +import org.apache.drill.exec.physical.impl.OutputMutator;
    +import org.apache.drill.exec.record.MaterializedField;
    +import org.apache.drill.exec.store.AbstractRecordReader;
    +import org.apache.drill.exec.store.pcap.decoder.Packet;
    +import org.apache.drill.exec.store.pcap.decoder.PacketDecoder;
    +import org.apache.drill.exec.store.pcap.dto.ColumnDto;
    +import org.apache.drill.exec.store.pcap.schema.PcapTypes;
    +import org.apache.drill.exec.store.pcap.schema.Schema;
    +import org.apache.drill.exec.vector.NullableBigIntVector;
    +import org.apache.drill.exec.vector.NullableIntVector;
    +import org.apache.drill.exec.vector.NullableTimeStampVector;
    +import org.apache.drill.exec.vector.NullableVarCharVector;
    +import org.apache.drill.exec.vector.ValueVector;
    +
    +import java.io.FileInputStream;
    +import java.io.IOException;
    +import java.io.InputStream;
    +import java.nio.ByteBuffer;
    +import java.util.List;
    +import java.util.Map;
    +
    +import static java.nio.charset.StandardCharsets.UTF_8;
    +import static org.apache.drill.exec.store.pcap.Utils.parseBytesToASCII;
    +
    +public class PcapRecordReader extends AbstractRecordReader {
    +
    +  private OutputMutator output;
    +
    +  private final PacketDecoder decoder;
    +  private ImmutableList<ProjectedColumnInfo> projectedCols;
    +
    +  private byte[] buffer = new byte[100000];
    +  private int offset = 0;
    +  private InputStream in;
    +  private int validBytes;
    +
    +  private static final Map<PcapTypes, MinorType> TYPES;
    +
    +  private static class ProjectedColumnInfo {
    +    ValueVector vv;
    +    ColumnDto pcapColumn;
    +  }
    +
    +  static {
    +    TYPES = ImmutableMap.<PcapTypes, TypeProtos.MinorType>builder()
    +        .put(PcapTypes.STRING, MinorType.VARCHAR)
    +        .put(PcapTypes.INTEGER, MinorType.INT)
    +        .put(PcapTypes.LONG, MinorType.BIGINT)
    +        .put(PcapTypes.TIMESTAMP, MinorType.TIMESTAMP)
    +        .build();
    +  }
    +
    +  public PcapRecordReader(final String inputPath,
    +                          final List<SchemaPath> projectedColumns) {
    +    try {
    +      this.in = new FileInputStream(inputPath);
    +      this.decoder = getPacketDecoder();
    +      validBytes = in.read(buffer);
    +    } catch (IOException e) {
    +      throw new RuntimeException("File " + inputPath + " not Found");
    +    }
    +    setColumns(projectedColumns);
    +  }
    +
    +  @Override
    +  public void setup(final OperatorContext context, final OutputMutator output) throws
ExecutionSetupException {
    +    this.output = output;
    +  }
    +
    +  @Override
    +  public int next() {
    +    projectedCols = getProjectedColsIfItNull();
    +    try {
    +      return parsePcapFilesAndPutItToTable();
    --- End diff --
    
    Drill has certain protocols that are not entirely obvious, but that are needed here. Each
call to `next()` produces a batch of rows. Drill requires that the batch contain no more than
64K rows. (Bad things happen if the batch has more.) So, the trick for a record reader is
to read up to some maximum number of records into each batch, return, and wait to be called
again for the next batch.
    
    Depending on the size of the data, 64K may be to large and 8K or some other number may
be better. Try to avoid causing any vector to be larger than 16 MB. (We're working on a project
to enforce this, and the result will make it much easier to write a record reader. Too bad
your project is just a bit too early to benefit.)
    
    If I read this code correctly, each call to `next()` returns just one record, until EOF
at which time it returns 0.
    
    You'll find that the rest of Drill will be *very* slow with a batch size of 1 record.
So, perhaps put this into a loop to read 8K or so rows into a single batch, then return the
total record count. You should see a nice execution speed-up.


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

Mime
View raw message