用户自定义 Sources & Sinks #

CREATE TABLE UserScores (name STRING, score INT)
WITH (
  'connector' = 'socket',
  'hostname' = 'localhost',
  'port' = '9999',
  'byte-delimiter' = '10',
  'format' = 'changelog-csv',
  'changelog-csv.column-delimiter' = '|'
);

SELECT name, SUM(score) FROM UserScores GROUP BY name;

> nc -lk 9999
INSERT|Alice|12
INSERT|Bob|5
DELETE|Alice|12
INSERT|Alice|18

import org.apache.flink.api.common.serialization.DeserializationSchema;
import org.apache.flink.configuration.ConfigOption;
import org.apache.flink.configuration.ConfigOptions;
import org.apache.flink.configuration.ReadableConfig;
import org.apache.flink.table.connector.format.DecodingFormat;
import org.apache.flink.table.connector.source.DynamicTableSource;
import org.apache.flink.table.data.RowData;
import org.apache.flink.table.factories.DeserializationFormatFactory;
import org.apache.flink.table.factories.DynamicTableSourceFactory;
import org.apache.flink.table.factories.FactoryUtil;
import org.apache.flink.table.types.DataType;

public class SocketDynamicTableFactory implements DynamicTableSourceFactory {

  // 定义所有配置项
  public static final ConfigOption<String> HOSTNAME = ConfigOptions.key("hostname")
    .stringType()
    .noDefaultValue();

  public static final ConfigOption<Integer> PORT = ConfigOptions.key("port")
    .intType()
    .noDefaultValue();

  public static final ConfigOption<Integer> BYTE_DELIMITER = ConfigOptions.key("byte-delimiter")
    .intType()
    .defaultValue(10); // 等同于 '\n'

  @Override
  public String factoryIdentifier() {
    return "socket"; // 用于匹配： `connector = '...'`
  }

  @Override
  public Set<ConfigOption<?>> requiredOptions() {
    final Set<ConfigOption<?>> options = new HashSet<>();
    options.add(HOSTNAME);
    options.add(PORT);
    options.add(FactoryUtil.FORMAT); // 解码的格式器使用预先定义的配置项
    return options;
  }

  @Override
  public Set<ConfigOption<?>> optionalOptions() {
    final Set<ConfigOption<?>> options = new HashSet<>();
    options.add(BYTE_DELIMITER);
    return options;
  }

  @Override
  public DynamicTableSource createDynamicTableSource(Context context) {
    // 使用提供的工具类或实现你自己的逻辑进行校验
    final FactoryUtil.TableFactoryHelper helper = FactoryUtil.createTableFactoryHelper(this, context);

    // 找到合适的解码器
    final DecodingFormat<DeserializationSchema<RowData>> decodingFormat = helper.discoverDecodingFormat(
      DeserializationFormatFactory.class,
      FactoryUtil.FORMAT);

    // 校验所有的配置项
    helper.validate();

    // 获取校验完的配置项
    final ReadableConfig options = helper.getOptions();
    final String hostname = options.get(HOSTNAME);
    final int port = options.get(PORT);
    final byte byteDelimiter = (byte) (int) options.get(BYTE_DELIMITER);

    // 从 catalog 中抽取要生产的数据类型 (除了需要计算的列) 
    final DataType producedDataType =
            context.getCatalogTable().getResolvedSchema().toPhysicalRowDataType();

    // 创建并返回动态表 source
    return new SocketDynamicTableSource(hostname, port, byteDelimiter, decodingFormat, producedDataType);
  }
}

import org.apache.flink.api.common.serialization.DeserializationSchema;
import org.apache.flink.configuration.ConfigOption;
import org.apache.flink.configuration.ConfigOptions;
import org.apache.flink.configuration.ReadableConfig;
import org.apache.flink.table.connector.format.DecodingFormat;
import org.apache.flink.table.data.RowData;
import org.apache.flink.table.factories.FactoryUtil;
import org.apache.flink.table.factories.DeserializationFormatFactory;
import org.apache.flink.table.factories.DynamicTableFactory;

public class ChangelogCsvFormatFactory implements DeserializationFormatFactory {

  // 定义所有配置项
  public static final ConfigOption<String> COLUMN_DELIMITER = ConfigOptions.key("column-delimiter")
    .stringType()
    .defaultValue("|");

  @Override
  public String factoryIdentifier() {
    return "changelog-csv";
  }

  @Override
  public Set<ConfigOption<?>> requiredOptions() {
    return Collections.emptySet();
  }

  @Override
  public Set<ConfigOption<?>> optionalOptions() {
    final Set<ConfigOption<?>> options = new HashSet<>();
    options.add(COLUMN_DELIMITER);
    return options;
  }

  @Override
  public DecodingFormat<DeserializationSchema<RowData>> createDecodingFormat(
      DynamicTableFactory.Context context,
      ReadableConfig formatOptions) {
    // 使用提供的工具类或实现你自己的逻辑进行校验
    FactoryUtil.validateFactoryOptions(this, formatOptions);

    // 获取校验完的配置项
    final String columnDelimiter = formatOptions.get(COLUMN_DELIMITER);

    // 创建并返回解码器
    return new ChangelogCsvFormat(columnDelimiter);
  }
}

import org.apache.flink.api.common.serialization.DeserializationSchema;
import org.apache.flink.legacy.table.connector.source.SourceFunctionProvider;
import org.apache.flink.streaming.api.functions.source.SourceFunction;
import org.apache.flink.table.connector.ChangelogMode;
import org.apache.flink.table.connector.format.DecodingFormat;
import org.apache.flink.table.connector.source.DynamicTableSource;
import org.apache.flink.table.connector.source.ScanTableSource;
import org.apache.flink.table.data.RowData;
import org.apache.flink.table.types.DataType;

public class SocketDynamicTableSource implements ScanTableSource {

  private final String hostname;
  private final int port;
  private final byte byteDelimiter;
  private final DecodingFormat<DeserializationSchema<RowData>> decodingFormat;
  private final DataType producedDataType;

  public SocketDynamicTableSource(
      String hostname,
      int port,
      byte byteDelimiter,
      DecodingFormat<DeserializationSchema<RowData>> decodingFormat,
      DataType producedDataType) {
    this.hostname = hostname;
    this.port = port;
    this.byteDelimiter = byteDelimiter;
    this.decodingFormat = decodingFormat;
    this.producedDataType = producedDataType;
  }

  @Override
  public ChangelogMode getChangelogMode() {
    // 在我们的例子中，由解码器来决定 changelog 支持的模式
    // 但是在 source 端指定也可以
    return decodingFormat.getChangelogMode();
  }

  @Override
  public ScanRuntimeProvider getScanRuntimeProvider(ScanContext runtimeProviderContext) {

    // 创建运行时类用于提交给集群

    final DeserializationSchema<RowData> deserializer = decodingFormat.createRuntimeDecoder(
      runtimeProviderContext,
      producedDataType);

    final SourceFunction<RowData> sourceFunction = new SocketSourceFunction(
      hostname,
      port,
      byteDelimiter,
      deserializer);

    return SourceFunctionProvider.of(sourceFunction, false);
  }

  @Override
  public DynamicTableSource copy() {
    return new SocketDynamicTableSource(hostname, port, byteDelimiter, decodingFormat, producedDataType);
  }

  @Override
  public String asSummaryString() {
    return "Socket Table Source";
  }
}

import org.apache.flink.api.common.serialization.DeserializationSchema;
import org.apache.flink.api.common.typeinfo.TypeInformation;
import org.apache.flink.table.connector.ChangelogMode;
import org.apache.flink.table.connector.format.DecodingFormat;
import org.apache.flink.table.connector.source.DynamicTableSource;
import org.apache.flink.table.connector.source.DynamicTableSource.DataStructureConverter;
import org.apache.flink.table.data.RowData;
import org.apache.flink.table.types.DataType;
import org.apache.flink.table.types.logical.LogicalType;
import org.apache.flink.types.RowKind;

public class ChangelogCsvFormat implements DecodingFormat<DeserializationSchema<RowData>> {

  private final String columnDelimiter;

  public ChangelogCsvFormat(String columnDelimiter) {
    this.columnDelimiter = columnDelimiter;
  }

  @Override
  @SuppressWarnings("unchecked")
  public DeserializationSchema<RowData> createRuntimeDecoder(
      DynamicTableSource.Context context,
      DataType producedDataType) {
    // 为 DeserializationSchema 创建类型信息
    final TypeInformation<RowData> producedTypeInfo = (TypeInformation<RowData>) context.createTypeInformation(
      producedDataType);

    // DeserializationSchema 中的大多数代码无法处理内部数据结构
    // 在最后为转换创建一个转换器
    final DataStructureConverter converter = context.createDataStructureConverter(producedDataType);

    // 在运行时，为解析过程提供逻辑类型
    final List<LogicalType> parsingTypes = producedDataType.getLogicalType().getChildren();

    // 创建运行时类
    return new ChangelogCsvDeserializer(parsingTypes, converter, producedTypeInfo, columnDelimiter);
  }

  @Override
  public ChangelogMode getChangelogMode() {
    // 支持处理 `INSERT`、`DELETE` 变更类型的数据。
    return ChangelogMode.newBuilder()
      .addContainedKind(RowKind.INSERT)
      .addContainedKind(RowKind.DELETE)
      .build();
  }
}

import org.apache.flink.api.common.serialization.DeserializationSchema;
import org.apache.flink.api.common.typeinfo.TypeInformation;
import org.apache.flink.table.connector.RuntimeConverter.Context;
import org.apache.flink.table.connector.source.DynamicTableSource.DataStructureConverter;
import org.apache.flink.table.data.RowData;
import org.apache.flink.table.types.logical.LogicalType;
import org.apache.flink.table.types.logical.LogicalTypeRoot;
import org.apache.flink.types.Row;
import org.apache.flink.types.RowKind;

public class ChangelogCsvDeserializer implements DeserializationSchema<RowData> {

  private final List<LogicalType> parsingTypes;
  private final DataStructureConverter converter;
  private final TypeInformation<RowData> producedTypeInfo;
  private final String columnDelimiter;

  public ChangelogCsvDeserializer(
      List<LogicalType> parsingTypes,
      DataStructureConverter converter,
      TypeInformation<RowData> producedTypeInfo,
      String columnDelimiter) {
    this.parsingTypes = parsingTypes;
    this.converter = converter;
    this.producedTypeInfo = producedTypeInfo;
    this.columnDelimiter = columnDelimiter;
  }

  @Override
  public TypeInformation<RowData> getProducedType() {
    // 为 Flink 的核心接口提供类型信息。
    return producedTypeInfo;
  }

  @Override
  public void open(InitializationContext context) {
    // 转化器必须要被开启。
    converter.open(Context.create(ChangelogCsvDeserializer.class.getClassLoader()));
  }

  @Override
  public RowData deserialize(byte[] message) {
    // 按列解析数据，其中一列是 changelog 标记
    final String[] columns = new String(message).split(Pattern.quote(columnDelimiter));
    final RowKind kind = RowKind.valueOf(columns[0]);
    final Row row = new Row(kind, parsingTypes.size());
    for (int i = 0; i < parsingTypes.size(); i++) {
      row.setField(i, parse(parsingTypes.get(i).getTypeRoot(), columns[i + 1]));
    }
    // 转换为内部数据结构
    return (RowData) converter.toInternal(row);
  }

  private static Object parse(LogicalTypeRoot root, String value) {
    switch (root) {
      case INTEGER:
        return Integer.parseInt(value);
      case VARCHAR:
        return value;
      default:
        throw new IllegalArgumentException();
    }
  }

  @Override
  public boolean isEndOfStream(RowData nextElement) {
    return false;
  }
}

import org.apache.flink.api.common.serialization.DeserializationSchema;
import org.apache.flink.api.common.typeinfo.TypeInformation;
import org.apache.flink.api.java.typeutils.ResultTypeQueryable;
import org.apache.flink.configuration.Configuration;
import org.apache.flink.streaming.api.functions.source.RichSourceFunction;
import org.apache.flink.table.data.RowData;

public class SocketSourceFunction extends RichSourceFunction<RowData> implements ResultTypeQueryable<RowData> {

  private final String hostname;
  private final int port;
  private final byte byteDelimiter;
  private final DeserializationSchema<RowData> deserializer;

  private volatile boolean isRunning = true;
  private Socket currentSocket;

  public SocketSourceFunction(String hostname, int port, byte byteDelimiter, DeserializationSchema<RowData> deserializer) {
    this.hostname = hostname;
    this.port = port;
    this.byteDelimiter = byteDelimiter;
    this.deserializer = deserializer;
  }

  @Override
  public TypeInformation<RowData> getProducedType() {
    return deserializer.getProducedType();
  }

  @Override
  public void open(OpenContext openContext) throws Exception {
    deserializer.open(() -> getRuntimeContext().getMetricGroup());
  }

  @Override
  public void run(SourceContext<RowData> ctx) throws Exception {
    while (isRunning) {
      // 持续从 socket 消费数据
      try (final Socket socket = new Socket()) {
        currentSocket = socket;
        socket.connect(new InetSocketAddress(hostname, port), 0);
        try (InputStream stream = socket.getInputStream()) {
          ByteArrayOutputStream buffer = new ByteArrayOutputStream();
          int b;
          while ((b = stream.read()) >= 0) {
            // 持续写入 buffer 直到遇到分隔符
            if (b != byteDelimiter) {
              buffer.write(b);
            }
            // 解码并处理记录
            else {
              ctx.collect(deserializer.deserialize(buffer.toByteArray()));
              buffer.reset();
            }
          }
        }
      } catch (Throwable t) {
        t.printStackTrace(); // 打印并继续
      }
      Thread.sleep(1000);
    }
  }

  @Override
  public void cancel() {
    isRunning = false;
    try {
      currentSocket.close();
    } catch (Throwable t) {
      // 忽略
    }
  }
}