问题导读:
1.代码从11升级到了12本文举例废弃了什么方法?
2.如何自定义的 sink类实现 DynamicTableSink 接口?
3.getChangelogMode设置什么?
一.变更简述:
最近代码从11升级到了12,由于此次版本变动,废弃了很多api,导致代码不兼容。如原来的ta b leEnv中注册registerTablesink和registerTableSource方法
- /** * Registers an external {@link TableSink} with given field names and types in this
复制代码
说明中要求通过 executeSql(ddl) 的方法注册一个表。
既然要通过ddl创建,那就要自定义一个redis 的connector。
flink自定的connector是借助 TableFactory spi发现注册的。
所以一共需要实现两个重要的类。
二 .关键类介绍
1.RedisDynamicTableSourceFactory类
自定义的facory类只需要实现 DynamicTableSinkFactory 接口即可。- import com.iqiyi.talos.engine.operators.sink.dynamic.RedisDynamicTableSink;
- import org.apache.flink.configuration.ConfigOption;
- import org.apache.flink.configuration.ReadableConfig;
- import org.apache.flink.table.connector.sink.DynamicTableSink;
- import org.apache.flink.table.factories.DynamicTableSinkFactory;
- import org.apache.flink.table.factories.FactoryUtil;
-
- import java.util.HashSet;
- import java.util.Set;
-
- /**
- * @ClassName RedisDynamicTableSourceFactory
- * @Description 自定义RedisSinkFactory
- * @Author zwc
- * @Date 2021-01-11 11:44
- * @Version 1.0
- **/
- public class RedisDynamicTableSourceFactory implements DynamicTableSinkFactory {
- // public static final ConfigOption<String> port = ConfigOptions.key("host").stringType().noDefaultValue();
-
-
- @Override
- public DynamicTableSink createDynamicTableSink(Context context) {
- final FactoryUtil.TableFactoryHelper helper = FactoryUtil.createTableFactoryHelper(this, context);
- helper.validate();
- ReadableConfig options = helper.getOptions();
- return new RedisDynamicTableSink(options);
- }
- //connector = "redis" 声明连接器名称
- @Override
- public String factoryIdentifier() {
- return "redis";
- }
-
- @Override
- public Set<ConfigOption<?>> requiredOptions() {
- Set<ConfigOption<?>> options = new HashSet();
- return options;
- }
-
- @Override
- public Set<ConfigOption<?>> optionalOptions() {
- Set<ConfigOption<?>> options = new HashSet();
- // options.add(port);
- return options;
- }
- }
复制代码
createDynamicTableSink: 校验ddl中 with(...) 附加的选项,并且从CatalogTable 初始化实例,将Options附加选项加载到上下文Context中。
factoryIdentifier: 连接器的名称
requiredOptions:必填参数
optionalOptions:可选参数
2.RedisDynamicTableSink类
自定义的 sink类实现 DynamicTableSink 接口
- import com.iqiyi.talos.engine.job.EngineContext;
- import com.iqiyi.talos.engine.operators.function.CollectionTableSinkFunction;
- import org.apache.flink.configuration.ReadableConfig;
- import org.apache.flink.table.connector.ChangelogMode;
- import org.apache.flink.table.connector.sink.DynamicTableSink;
- import org.apache.flink.table.connector.sink.SinkFunctionProvider;
-
- /**
- * @ClassName RedisDynamicTableSink
- * @Description TODO
- * @Author zwc
- * @Date 2021-01-11 15:41
- * @Version 1.0
- **/
- public class RedisDynamicTableSink implements DynamicTableSink {
-
- private ReadableConfig options;
-
- private EngineContext ctx;
-
- public RedisDynamicTableSink(ReadableConfig options) {
- this.options = options;
- ctx = EngineContext.getContext();
- }
-
-
- @Override
- public ChangelogMode getChangelogMode(ChangelogMode requestedMode) {
- return ChangelogMode.insertOnly();
- }
-
- @Override
- public SinkRuntimeProvider getSinkRuntimeProvider(Context context) {
-
- CollectionTableSinkFunction collectionTableSinkFunction = new CollectionTableSinkFunction(ctx.getDeploySite(), ctx.getSimpifiedName());
- return SinkFunctionProvider.of(collectionTableSinkFunction);
- }
-
- @Override
- public DynamicTableSink copy() {
- return new RedisDynamicTableSink(this.options);
- }
-
- @Override
- public String asSummaryString() {
- return "my_redis_sink";
- }
- }
复制代码
getChangelogMode: 设置sink 是 insert only模式(目前只支持这种模式)
getSinkRuntimeProvider: 这里就是执行sink的具体逻辑了。这里可以直接使用flink 提供的redis-connector
- <dependency>
- <groupId>org.apache.bahir</groupId>
- <artifactId>flink-connector-redis_2.11</artifactId>
- <version>1.0</version>
- </dependency>
复制代码
- RedisMapper<RowData> stringRedisMapper = new RedisMapper<RowData>() {
- //返回对应Redis命令
- @Override
- public RedisCommandDescription getCommandDescription() {
- return new RedisCommandDescription(RedisCommand.SET);
- }
- //从数据中获取对应Key
- @Override
- public String getKeyFromData(RowData rowData) {
- StringData string = rowData.getString(0);
- return string;
- }
- //从数据中获取对应Value
- @Override
- public String getValueFromData(RowData rowData) {
- String s = rowData.toString();
- return s;
复制代码
由于此处我需要自定义redis数据结构,所以自己实现了RichFunction
- @Override
- public SinkRuntimeProvider getSinkRuntimeProvider(Context context) {
-
- CollectionTableSinkFunction collectionTableSinkFunction = new CollectionTableSinkFunction(ctx.getDeploySite(), ctx.getSimpifiedName());
- return SinkFunctionProvider.of(collectionTableSinkFunction);
- }
复制代码
- /**
- * @ClassName CollectionTableSinkFunction
- * @Description TODO
- * @Author zwc
- * @Date 2021-01-12 16:51
- * @Version 1.0
- **/
-
- import com.alibaba.fastjson.JSON;
- import com.iqiyi.talos.common.JedisClient;
- import com.iqiyi.talos.engine.enums.DeploySite;
- import org.apache.flink.configuration.Configuration;
- import org.apache.flink.streaming.api.functions.sink.RichSinkFunction;
- import org.apache.flink.table.data.RowData;
- import org.slf4j.Logger;
- import org.slf4j.LoggerFactory;
-
- import java.util.ArrayList;
- import java.util.List;
- import java.util.Map;
- import java.util.concurrent.ConcurrentHashMap;
-
- /**
- * 这里定义了当接收到一条数据时,该如何 sink 的具体逻辑
- */
- public class CollectionTableSinkFunction extends RichSinkFunction<RowData> {
- private static Logger LOG = LoggerFactory.getLogger(CollectionTableSink.CollectionTableSinkFunction.class);
- private static Map<String, List<String>> map = new ConcurrentHashMap<>();
- private DeploySite deploySite;
- private String jobName;
- private static final Object lock = new Object();
-
- public CollectionTableSinkFunction(DeploySite deploySite, String jobName) {
- this.deploySite = deploySite;
- this.jobName = jobName;
- }
-
- @Override
- public void open(Configuration parameters) throws Exception {
- super.open(parameters);
- LOG.info("OPEN!");
- }
-
- @Override
- public void invoke(RowData row, Context context) throws Exception {
- String operatorName = row.getString(0).toString();
- String value = row.getString(1).toString();
- //operatorName source 1
-
- map.compute(operatorName, (k,v) -> {
- if(v==null) {
- v = new ArrayList<>();
- v.add(value);
- } else {
- v.add(value);
- }
- return v;
- });
- LOG.info("put to collectionMap. [{},{}] ", operatorName, value);
- LOG.info("mapSize:" + map.size());
- }
-
- public Map<String,List<String>> getContent() {
- return map;
- }
-
- @Override
- public void close() {
- synchronized (lock) {
- Map<String,List<String>> map = getContent();
- String key = "TalosMockJob_" + jobName;
- String value = JSON.toJSONString(map);
- LOG.info("Send Mock result to redis. key:{}, value:{}", key, value);
- long ttl = 24 * 3600 * 1000;
- try {
- JedisClient.get(deploySite.name()).setValue(key, value, ttl);
- } catch (Exception e) {
-
- }
- }
- }
- }
复制代码
3.最后一步spi
把RedisDynamicTableSourceFactory类包路径填进去即可
4.使用
- ctx.getTableEnv().executeSql("create table MockJob_Data_Table (\n" +
- " operatorName STRING," +
- " data STRING" +
- ") WITH (\n" +
- " 'connector' = 'redis'\n" +
- ")");
复制代码
原文链接: https://www.jianshu.com/p/48206a78425d
|