假设有一张下面这样的 MOR 类型的 hudi 表:
CREATE TABLE t_test_001 (
ds BIGINT COMMENT 'ds',
ut BIGINT COMMENT 'ut',
pk BIGINT COMMENT 'pk',
a0 BIGINT COMMENT 'a0',
a1 BIGINT COMMENT 'a1',
a2 BIGINT COMMENT 'a2'
) USING hudi
则在 drop 该表时,正确的 drop 方法为:
DROP TABLE t_test_001_rt;
DROP TABLE t_test_001_ro;
DROP TABLE t_test_001;
即带尾巴的要放到最后,否则带尾巴的表 drop 不掉:
org.apache.hudi.exception.HoodieIOException: Could not check if hdfs:///user/root/test.db/t_test_001 is a valid table
Caused by: java.io.FileNotFoundException: File does not exist: hdfs:///user/root/test.db/t_test_001/.hoodie
如果只 drop 不带尾巴的表,则使用时会报错:
Caused by: java.sql.SQLException: java.util.concurrent.ExecutionException: java.lang.RuntimeException: org.apache.hudi.exception.HoodieException: Could not sync using the meta sync class org.apache.hudi.hive.HiveSyncTool
... 25 more
Caused by: org.apache.hudi.hive.HoodieHiveSyncException: Could not convert field Type from STRING to bigint for field a2
... 36 more