paimon手动清除分区数据

环境:paimon分区可以设置为自动失效,但官网也提供了手动方案,paimon 0.5版本只支持flink,本测试案例 flink 1.17.1 paimon 0.5,hive 3.1.2

案例

启动yarn-session

flink client操作

[root@ks2p-hadoop06 flink-1.17.1]# ./bin/yarn-session.sh 
# 连接yarn session
[root@ks2p-hadoop06 bin]# ./sql-client.sh

Flink SQL> CREATE CATALOG paimon_hive WITH (
>     'type' = 'paimon',
>     'metastore' = 'hive',
>     'uri' = 'thrift://xx.xx.xx.xx:9083',
>     'warehouse' = 'hdfs:///data/hive/warehouse/paimon',
>     'default-database'='test'
> );

# 改为批模式执行
Flink SQL> SET 'execution.runtime-mode' = 'batch';

Flink SQL> INSERT OVERWRITE test.student2 /*+ OPTIONS('dynamic-partition-overwrite'='false') */ PARTITION (dt = '2023-10-09') SELECT id,name,age,birthday  FROM test.student2 WHERE false;