Trun On ACID(Detlete & Update) On Hive

序言

默认我们搭建起来的Hive或者SparkSql都是不支持事务的.需要相应的配置才能开启Hive的事务.

同时Hive的Delete和Update也是需要先开启ACID才能支持的cuiyaonan2000@163.com.

参考地址:

1. LanguageManual DML - Apache Hive - Apache Software Foundation
2. Hive Transactions - Apache Hive - Apache Software Foundation

 

 配置文件

还是针对hive-site.xml进项额外的配置即可

hive.support.concurrencytruehive.enforce.bucketingtruehive.exec.dynamic.partition.modenonstricthive.txn.managerorg.apache.hadoop.hive.ql.lockmgr.DbTxnManagerhive.compactor.initiator.ontruehive.compactor.worker.threads1

数据库

如果你是在已经初始化的metastore上启用ACID则需要执行hive目录/soft/hadoop/apache-hive-3.1.2-bin/scripts/metastore/upgrade/mysql 下找到如下的文件在mysql的数据库中执行cuiyaonan2000@163.com

 

如果是还没初始化数据库则,在配置完hive-site.xml后,直接执行初始化数据库就行了.命令如下:

schematool -dbType mysql -initSchema

Table的相关设置

 通过官网和网友们的翻译我们了解到table必须满足如下的要求才能支持事务和delete与update.

1. 
2.  
3. 

翻译如下

1. 表的存储格式必须是ORC（STORED AS ORC）；
2. 表必须进行分桶（CLUSTERED BY (col_name, col_name, ...)  INTO num_buckets BUCKETS）；
3. Table property中参数transactional必须设定为True（tblproperties('transactional'='true')）；

官网的示例:

CREATE TABLE table_name (
  id                int,
  name              string
)
CLUSTERED BY (id) INTO 2 BUCKETS STORED AS ORC
TBLPROPERTIES ("transactional"="true",
  "compactor.mapreduce.map.memory.mb"="2048",     -- specify compaction map job properties
  "compactorthreshold.hive.compactor.delta.num.threshold"="4",  -- trigger minor compaction if there are more than 4 delta directories
  "compactorthreshold.hive.compactor.delta.pct.threshold"="0.5" -- trigger major compaction if the ratio of size of delta files to
                                                                   -- size of base files is greater than 50%
);