Java Group类代码示例

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
@Override
public int run(String[] args) throws Exception {
    if (args.length != 2) {
        System.err.printf("Usage: %s [generic options] <input> <output>\n",
                getClass().getSimpleName());
        ToolRunner.printGenericCommandUsage(System.err);
        return -1;
    }

    Job job = new Job(getConf(), "Text to Parquet");
    job.setJarByClass(getClass());
    FileInputFormat.addInputPath(job, new Path(args[0]));
    FileOutputFormat.setOutputPath(job, new Path(args[1]));
    job.setMapperClass(TextToParquetMapper.class);
    job.setNumReduceTasks(0);
    job.setOutputFormatClass(AvroParquetOutputFormat.class);
    AvroParquetOutputFormat.setSchema(job, SCHEMA);
    job.setOutputKeyClass(Void.class);
    job.setOutputValueClass(Group.class);
    return job.waitForCompletion(true) ? 0 : 1;
}
 

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
public static void main(String[] args) throws IOException {

    SimpleGroupFactory sgf = new SimpleGroupFactory(simpleSchema);
    GroupFactory gf = new SimpleGroupFactory(complexSchema);
    SimpleGroupFactory sngf = new SimpleGroupFactory(simpleNullableSchema);
    GroupFactory ngf = new SimpleGroupFactory(complexNullableSchema);

    ParquetWriter<Group> simpleWriter = initWriter(simpleSchema, "drill/parquet_test_file_simple");
    ParquetWriter<Group> complexWriter = initWriter(complexSchema, "drill/parquet_test_file_complex");
    ParquetWriter<Group> simpleNullableWriter = initWriter(simpleNullableSchema, "drill/parquet_test_file_simple_nullable");
    ParquetWriter<Group> complexNullableWriter = initWriter(complexNullableSchema, "drill/parquet_test_file_complex_nullable");

    ParquetSimpleTestFileGenerator.writeSimpleValues(sgf, simpleWriter, false);
    ParquetSimpleTestFileGenerator.writeSimpleValues(sngf, simpleNullableWriter, true);
    ParquetSimpleTestFileGenerator.writeComplexValues(gf, complexWriter, false);
    ParquetSimpleTestFileGenerator.writeComplexValues(ngf, complexNullableWriter, true);

    simpleWriter.close();
    complexWriter.close();
    simpleNullableWriter.close();
    complexNullableWriter.close();

  }
 

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
static File generateEmptyWithSchema(File parentDir, String filename) throws IOException {
  File f = new File(parentDir, filename);

  Configuration conf = new Configuration();
  MessageType schema = parseMessageType(
      "message test { "
          + "required int32 int32_field; "
          + "required int64 int64_field; "
          + "required float float_field; "
          + "required double double_field; "
          + "required int64 timestamp_field (TIMESTAMP_MILLIS);"
          + "} ");
  GroupWriteSupport.setSchema(schema, conf);
  SimpleGroupFactory fact = new SimpleGroupFactory(schema);
  ParquetWriter<Group> writer = new ParquetWriter<Group>(new Path(f.getPath()), new GroupWriteSupport(),
      UNCOMPRESSED, 1024, 1024, 512, false, false, ParquetProperties.WriterVersion.PARQUET_2_0, conf);
  writer.close();

  return f;
}
 

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
static File generateSparseParquetFile(File parentDir, String filename, int nrows) throws IOException {
  File f = new File(parentDir, filename);

  Configuration conf = new Configuration();
  MessageType schema = parseMessageType(
          "message test { optional int32 int32_field; optional binary string_field (UTF8); required int32 row; optional int32 int32_field2; } ");
  GroupWriteSupport.setSchema(schema, conf);
  SimpleGroupFactory fact = new SimpleGroupFactory(schema);
  ParquetWriter<Group> writer = new ParquetWriter<Group>(new Path(f.getPath()), new GroupWriteSupport(),
          UNCOMPRESSED, 1024, 1024, 512, true, false, ParquetProperties.WriterVersion.PARQUET_2_0, conf);
  try {
    for (int i = 0; i < nrows; i++) {
      Group g = fact.newGroup();
      if (i % 10 == 0) { g = g.append("int32_field", i); }
      if (i % 10 == 0) { g = g.append("string_field", "CAT_" + (i % 10)); }
      if (i % 10 == 0) { g = g.append("int32_field2", i); }
      writer.write(g.append("row", i));
    }
  } finally {
    writer.close();
  }
  return f;
}
 

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
static File generateParquetFileWithNullCharacters(File parentDir, String filename, int nrows) throws IOException {
  File f = new File(parentDir, filename);

  Configuration conf = new Configuration();
  MessageType schema = parseMessageType(
          "message test { optional binary cat_field (UTF8); } ");
  GroupWriteSupport.setSchema(schema, conf);
  SimpleGroupFactory fact = new SimpleGroupFactory(schema);
  ParquetWriter<Group> writer = new ParquetWriter<Group>(new Path(f.getPath()), new GroupWriteSupport(),
          UNCOMPRESSED, 1024, 1024, 512, true, false, ParquetProperties.WriterVersion.PARQUET_2_0, conf);
  try {
    for (int i = 0; i < nrows; i++) {
      Group g = fact.newGroup();
      String value = i == 66 ? "CAT_0_weird\0" : "CAT_" + (i % 10);
      writer.write(g.append("cat_field", value));
    }
  } finally {
    writer.close();
  }
  return f;
}
 

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
private void read(Path parquetFile, int nRows, Blackhole blackhole) throws IOException
{
  ParquetReader<Group> reader = ParquetReader.builder(new GroupReadSupport(), parquetFile).withConf(configuration).build();
  for (int i = 0; i < nRows; i++) {
    Group group = reader.read();
    blackhole.consume(group.getBinary("binary_field", 0));
    blackhole.consume(group.getInteger("int32_field", 0));
    blackhole.consume(group.getLong("int64_field", 0));
    blackhole.consume(group.getBoolean("boolean_field", 0));
    blackhole.consume(group.getFloat("float_field", 0));
    blackhole.consume(group.getDouble("double_field", 0));
    blackhole.consume(group.getBinary("flba_field", 0));
    blackhole.consume(group.getInt96("int96_field", 0));
  }
  reader.close();
}
 

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
public String toString(String indent) {
  String result = "";
  int i = 0;
  for (Type field : schema.getFields()) {
    String name = field.getName();
    List<Object> values = data[i];
    ++i;
    if (values != null) {
      if (values.size() > 0) {
        for (Object value : values) {
          result += indent + name;
          if (value == null) {
            result += ": NULL\n";
          } else if (value instanceof Group) {
            result += "\n" + ((SimpleGroup)value).toString(indent+"  ");
          } else {
            result += ": " + value.toString() + "\n";
          }
        }
      }
    }
  }
  return result;
}
 

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
@Test
public void testReadUsingProjectedSchema(){
  MessageType orginalSchema = new MessageType("schema",
          new PrimitiveType(REQUIRED, INT32, "a"),
          new PrimitiveType(REQUIRED, INT32, "b")
  );
  MessageType projectedSchema = new MessageType("schema",
          new PrimitiveType(OPTIONAL, INT32, "b")
  );
  MemPageStore store = new MemPageStore(1);
  SimpleGroupFactory groupFactory = new SimpleGroupFactory(orginalSchema);
  writeGroups(orginalSchema, store, groupFactory.newGroup().append("a", 1).append("b", 2));

  {
    List<Group> groups = new ArrayList<Group>();
    groups.addAll(readGroups(store, orginalSchema, projectedSchema, 1));
    Object[][] expected = {
            {2},
    };
    validateGroups(groups, expected);
  }
}
 

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
private void validateGroups(List<Group> groups1, Object[][] e1) {
  Iterator<Group> i1 = groups1.iterator();
  for (int i = 0; i < e1.length; i++) {
    Object[] objects = e1[i];
    Group next = i1.next();
    for (int j = 0; j < objects.length; j++) {
      Object object = objects[j];
      if (object == null) {
        assertEquals(0, next.getFieldRepetitionCount(j));
      } else {
        assertEquals("looking for r[" + i + "][" + j + "][0]=" + object, 1, next.getFieldRepetitionCount(j));
        assertEquals(object, next.getInteger(j, 0));
      }
    }
  }
}
 

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
@Test
public void testOneOfEach() {
  MessageType oneOfEachSchema = MessageTypeParser.parseMessageType(oneOfEach);
  GroupFactory gf = new SimpleGroupFactory(oneOfEachSchema);
  Group g1 = gf.newGroup()
      .append("a", 1l)
      .append("b", 2)
      .append("c", 3.0f)
      .append("d", 4.0d)
      .append("e", true)
      .append("f", Binary.fromString("6"))
      .append("g", new NanoTime(1234, System.currentTimeMillis() * 1000))
      .append("h", Binary.fromString("abc"));

  testSchema(oneOfEachSchema, Arrays.asList(g1));
}
 

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
@Test
public void testFilterOnInteger() {
  MessageColumnIO columnIO =  new ColumnIOFactory(true).getColumnIO(schema);
  MemPageStore memPageStore = writeTestRecords(columnIO, 1);

  // Get first record
  RecordMaterializer<Group> recordConverter = new GroupRecordConverter(schema);
  RecordReaderImplementation<Group> recordReader = (RecordReaderImplementation<Group>)
      columnIO.getRecordReader(memPageStore, recordConverter, FilterCompat.get(column("DocId", equalTo(10l))));

  readOne(recordReader, "r2 filtered out", r1);

  // Get second record
  recordReader = (RecordReaderImplementation<Group>)
      columnIO.getRecordReader(memPageStore, recordConverter,
          FilterCompat.get(column("DocId", equalTo(20l))));

  readOne(recordReader, "r1 filtered out", r2);

}
 

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
@Test
public void testApplyFunctionFilterOnLong() {
  MessageColumnIO columnIO =  new ColumnIOFactory(true).getColumnIO(schema);
  MemPageStore memPageStore = writeTestRecords(columnIO, 1);

  // Get first record
  RecordMaterializer<Group> recordConverter = new GroupRecordConverter(schema);
  RecordReaderImplementation<Group> recordReader = (RecordReaderImplementation<Group>)
      columnIO.getRecordReader(memPageStore, recordConverter,
          FilterCompat.get(column("DocId", equalTo(10l))));

  readOne(recordReader, "r2 filtered out", r1);

  // Get second record
  recordReader = (RecordReaderImplementation<Group>)
      columnIO.getRecordReader(memPageStore, recordConverter,
          FilterCompat.get(column("DocId", applyFunctionToLong(new LongGreaterThan15Predicate()))));

  readOne(recordReader, "r1 filtered out", r2);
}
 

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
@Test
public void testPaged() {
  MessageColumnIO columnIO =  new ColumnIOFactory(true).getColumnIO(schema);
  MemPageStore memPageStore = writeTestRecords(columnIO, 6);

  RecordMaterializer<Group> recordConverter = new GroupRecordConverter(schema);
  RecordReaderImplementation<Group> recordReader = (RecordReaderImplementation<Group>)
      columnIO.getRecordReader(memPageStore, recordConverter,
          FilterCompat.get(page(4, 4)));

  List<Group> all = readAll(recordReader);
  assertEquals("expecting records " + all, 4, all.size());
  for (int i = 0; i < all.size(); i++) {
    assertEquals("expecting record", (i%2 == 0 ? r2 : r1).toString(), all.get(i).toString());
  }
}
 

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
@Test
public void testFilteredAndPaged() {
  MessageColumnIO columnIO =  new ColumnIOFactory(true).getColumnIO(schema);
  MemPageStore memPageStore = writeTestRecords(columnIO, 8);

  RecordMaterializer<Group> recordConverter = new GroupRecordConverter(schema);
  RecordReaderImplementation<Group> recordReader = (RecordReaderImplementation<Group>)
      columnIO.getRecordReader(memPageStore, recordConverter,
          FilterCompat.get(and(column("DocId", equalTo(10l)), page(2, 4))));

  List<Group> all = readAll(recordReader);
  assertEquals("expecting 4 records " + all, 4, all.size());
  for (int i = 0; i < all.size(); i++) {
    assertEquals("expecting record1", r1.toString(), all.get(i).toString());
  }

}
 

import org.apache.parquet.example.data.Group; //导入依赖的package包/类
@Test
public void testFilteredOrPaged() {
  MessageColumnIO columnIO =  new ColumnIOFactory(true).getColumnIO(schema);
  MemPageStore memPageStore = writeTestRecords(columnIO, 8);

  RecordMaterializer<Group> recordConverter = new GroupRecordConverter(schema);
  RecordReaderImplementation<Group> recordReader = (RecordReaderImplementation<Group>)
      columnIO.getRecordReader(memPageStore, recordConverter,
          FilterCompat.get(or(column("DocId", equalTo(10l)),
              column("DocId", equalTo(20l)))));

  List<Group> all = readAll(recordReader);
  assertEquals("expecting 8 records " + all, 16, all.size());
  for (int i = 0; i < all.size () / 2; i++) {
    assertEquals("expecting record1", r1.toString(), all.get(2 * i).toString());
    assertEquals("expecting record2", r2.toString(), all.get(2 * i + 1).toString());
  }
}