删除hive分区外部表但保留分区

use*_*851 3 hadoop hive metadata hdfs

使用外部配置单元表时,有没有办法可以删除目录中的数据,但通过查询保留分区.请注意我不想放弃桌子并重新创建它.我只是想清空底层文件夹并重新开始一个进程.我的表很大,按年,月,日和小时划分分区,手动重新创建分区需要花费大量时间.

谢谢

Dav*_*itz 6

truncate table ...删除所有数据.
truncate table partition (...)删除特定分区的数据.

保留目录结构.


外部表应首先转换为manged,.eg

alter table t set tblproperties('EXTERNAL'='FALSE');
Run Code Online (Sandbox Code Playgroud)

当我们完成后,我们可以将其转换回来

alter table t set tblproperties('EXTERNAL'='TRUE');
Run Code Online (Sandbox Code Playgroud)

演示

create table t (i int) partitioned by (x char(1));
set hive.exec.dynamic.partition.mode=nonstrict;
insert into t partition (x) values (1,'A'),(2,'B'),(3,'C');
alter table t set tblproperties('EXTERNAL'='TRUE');
Run Code Online (Sandbox Code Playgroud)
select * from t;
Run Code Online (Sandbox Code Playgroud)
+-----+-----+
| t.i | t.x |
+-----+-----+
|   1 | A   |
|   2 | B   |
|   3 | C   |
+-----+-----+
Run Code Online (Sandbox Code Playgroud)
dfs -ls -R /user/hive/warehouse/t;
Run Code Online (Sandbox Code Playgroud)
drwxrwxrwx   - cloudera supergroup          0 2017-03-28 11:40 /user/hive/warehouse/t/x=A
-rwxrwxrwx   1 cloudera supergroup          2 2017-03-28 11:40 /user/hive/warehouse/t/x=A/000000_0
drwxrwxrwx   - cloudera supergroup          0 2017-03-28 11:40 /user/hive/warehouse/t/x=B
-rwxrwxrwx   1 cloudera supergroup          2 2017-03-28 11:40 /user/hive/warehouse/t/x=B/000000_0
drwxrwxrwx   - cloudera supergroup          0 2017-03-28 11:40 /user/hive/warehouse/t/x=C
-rwxrwxrwx   1 cloudera supergroup          2 2017-03-28 11:40 /user/hive/warehouse/t/x=C/000000_0
Run Code Online (Sandbox Code Playgroud)
truncate table t partition (x='B');
Run Code Online (Sandbox Code Playgroud)

FAILED:SemanticException [错误10146]:无法截断非托管表t.

alter table t set tblproperties('EXTERNAL'='FALSE');
Run Code Online (Sandbox Code Playgroud)
truncate table t partition (x='B');
Run Code Online (Sandbox Code Playgroud)
select * from t;
Run Code Online (Sandbox Code Playgroud)
+-----+-----+
| t.i | t.x |
+-----+-----+
|   1 | A   |
|   3 | C   |
+-----+-----+
Run Code Online (Sandbox Code Playgroud)
dfs -ls -R /user/hive/warehouse/t;
Run Code Online (Sandbox Code Playgroud)
drwxrwxrwx   - cloudera supergroup          0 2017-03-28 11:40 /user/hive/warehouse/t/x=A
-rwxrwxrwx   1 cloudera supergroup          2 2017-03-28 11:40 /user/hive/warehouse/t/x=A/000000_0
drwxrwxrwx   - cloudera supergroup          0 2017-03-28 11:42 /user/hive/warehouse/t/x=B
drwxrwxrwx   - cloudera supergroup          0 2017-03-28 11:40 /user/hive/warehouse/t/x=C
-rwxrwxrwx   1 cloudera supergroup          2 2017-03-28 11:40 /user/hive/warehouse/t/x=C/000000_0
Run Code Online (Sandbox Code Playgroud)
truncate table t;
Run Code Online (Sandbox Code Playgroud)
+-----+-----+
| t.i | t.x |
+-----+-----+
Run Code Online (Sandbox Code Playgroud)
dfs -ls -R /user/hive/warehouse/t;
Run Code Online (Sandbox Code Playgroud)
drwxrwxrwx   - cloudera supergroup          0 2017-03-28 11:43 /user/hive/warehouse/t/x=A
drwxrwxrwx   - cloudera supergroup          0 2017-03-28 11:43 /user/hive/warehouse/t/x=B
drwxrwxrwx   - cloudera supergroup          0 2017-03-28 11:43 /user/hive/warehouse/t/x=C
Run Code Online (Sandbox Code Playgroud)
alter table t set tblproperties('EXTERNAL'='TRUE');
Run Code Online (Sandbox Code Playgroud)