我正在使用 flink 1.12.0。尝试将数据流转换为表 A 并在表 A 上运行 sql 查询以在窗口上聚合,如下所示。我使用 f2 列作为其时间戳数据类型字段。
EnvironmentSettings fsSettings = EnvironmentSettings.newInstance().useBlinkPlanner().inStreamingMode().build();
final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
StreamTableEnvironment tEnv = StreamTableEnvironment.create(env, fsSettings);
Properties props = new Properties();
props.setProperty("bootstrap.servers", BOOTSTRAP_SERVER);
props.setProperty("schema.registry.url", xxx);
props.setProperty("group.id", "test");
props.setProperty("value.deserializer", KafkaAvroDeserializer.class.getName());
props.put("client.id", "flink-kafka-example");
FlinkKafkaConsumer<PlaybackListening> kafkaConsumer = new FlinkKafkaConsumer<>(
"test-topic",
ConfluentRegistryAvroDeserializationSchema.forSpecific(
Avrotest.class, prodSchemaRegistryURL),
props);
DataStreamSource<Avrotest> stream =
env.addSource(kafkaConsumer);
Table tableA = tEnv.fromDataStream(stream, $("f0"), $("f1"),$("f2"));
Table result =
tEnv.sqlQuery("SELECT f0, sum(f1),f2 FROM "
+ tableA + " GROUP BY TUMBLE(f2, INTERVAL '1' HOUR) ,f1" );
tEnv.toAppendStream(result,user.class).print();
env.execute("Flink kafka test");
}
当我执行上面的代码时,我得到
线程“主”org.apache.flink.table.api.TableException 中的异常:窗口聚合只能在时间属性列上定义,但遇到 TIMESTAMP(6)。在 org.apache.flink.table.planner.plan.rules.logical.StreamLogicalWindowAggregateRule.getInAggregateGroupExpression(StreamLogicalWindowAggregateRule.scala:50) 在 org.apache.flink.table.planner.plan.rules.logical.LogicalWindowAggregateRuleBase.onMatch(LogicalWindowAggregateRuleBase. scala:81) 在 org.apache.calcite.plan.AbstractRelOptPlanner.fireRule(AbstractRelOptPlanner.java:333) 在 org.apache.calcite.plan.hep.HepPlanner.applyRule(HepPlanner.java:542) 在 org.apache.calcite .plan.hep.HepPlanner.applyRules(HepPlanner.java:407) 在 org.apache.calcite.plan.hep.HepPlanner.executeInstruction(HepPlanner.java:243) 在 org.apache.calcite.plan.hep.HepInstruction$RuleInstance .