【flink SQL基础｜概念】flink sql之时态表：join时要使用proctime作为时态表的time

flink版本1.12.5

1. 时态表

看下官网的定义

时态表（Temporal Table）代表基于表的（参数化）视图概念，该表记录变更历史，该视图返回表在某个特定时间点的内容。

“时”代表一个时间点，那“态”就代表着这一时刻表的状态，那时态表可以简单认为：这一时刻查询这张表的结果（或是快照）。举个例子：mysql表中11点更新了一条数据，那11点前后查询的结果是不同的，所以时态表（Temporal Table）是一张随时间变化的表。

Flink 使用 SQL:2011 中提出的 FOR SYSTEM_TIME AS OF 的 SQL 语法查询时态表。目前，时态表只能在 join 中使用。

接下来我们看一个join的例子

2. join的例子

CREATE TABLE `send_table01` (
   `id` INTEGER,
   `address` STRING,
   `cloud_wise_proc_time` AS `proctime`(),
   `cloud_wise_event_time` TIMESTAMP(3),
   WATERMARK FOR `cloud_wise_event_time` AS `cloud_wise_event_time` - INTERVAL '1' MINUTE
 ) WITH (
   'connector' = 'kafka',
   'topic' = 'topic_source_004',
   'properties.bootstrap.servers' = '10.0.16.44:18108',
   'properties.max.poll.records' = '5000',
   'properties.group.id' = 'wrwrwrw',
   'format' = 'json',
   'json.ignore-parse-errors' = 'true',
   'scan.startup.mode' = 'latest-offset'
 );

CREATE TABLE `send_table02` (
   `name` STRING,
   `id` INTEGER,
   PRIMARY KEY (`id`) NOT ENFORCED
 ) WITH (
   'connector' = 'jdbc',
   'driver' = 'ru.yandex.clickhouse.ClickHouseDriver',
   'url' = 'jdbc:clickhouse://10.101.1.234:18100/docp?socket_timeout=900000',
   'username' = 'default'
   'table-name' = 'pop'
 );

CREATE TABLE `out_table` (
   `name` STRING NOT NULL
 ) WITH (
   'connector' = 'kafka',
   'topic' = 'topic_out_60',
   'properties.bootstrap.servers' = '10.0.16.44:18108',
   'properties.acks' = '1',
   'properties.retries' = '0',
   'properties.batch.size' = '16348',
   'properties.linger.ms' = '0',
   'properties.buffer.memory' = '33554432',
   'json.ignore-parse-errors' = 'true',
   'format' = 'json'
 );

INSERT INTO `out_table`
 (SELECT `b`.`name` AS `name`
 FROM `send_table01` AS `a`
 INNER JOIN `send_table02` FOR SYSTEM_TIME AS OF `a`.`cloud_wise_proc_time` AS `b` ON `a`.`id` = `b`.`id`);

注意：

INNER JOIN `send_table02` FOR SYSTEM_TIME AS OF `a`.`cloud_wise_proc_time` AS `b`

cloud_wise_proc_time字段要为处理时间，对于上述例子，当flink消费kafka的数据然后通过关联字段查询mysql时，flink 将cloud_wise_proc_time作为时态表的”时”去查询。也就是：kafka来一条数据的时刻作为查询mysql（时态表）的“时”。

如果选择kafka数据中事件时间作为时态表的“时”，假设数据中有事件时间为去年某刻，此时flink可能会报错，因为去年这张mysql表还不存在。

所以，针对于这样的维表join场景，当flink报相关错误时，可以考虑是否是时间类型用错了

Event-Time Temporal Table Join requires both primary key and row time attribute in versioned table, but no row time attribute can be found.
org.apache.flink.table.api.ValidationException: Event-Time Temporal Table Join requires both primary key and row time attribute in versioned table, but no row time attribute can be found.
	at org.apache.flink.table.planner.plan.rules.logical.LogicalCorrelateToJoinFromGeneralTemporalTableRule.onMatch(LogicalCorrelateToJoinFromTemporalTableRule.scala:285) ~[flink-table-blink_2.11-1.12.5.jar:1.12.5]
	at org.apache.calcite.plan.AbstractRelOptPlanner.fireRule(AbstractRelOptPlanner.java:333) ~[flink-table-blink_2.11-1.12.5.jar:1.12.5]
	at org.apache.calcite.plan.hep.HepPlanner.applyRule(HepPlanner.java:542) ~[flink-table-blink_2.11-1.12.5.jar:1.12.5]
	at org.apache.calcite.plan.hep.HepPlanner.applyRules(