有時我們需要批量刪除一些hbase中符合某些條件的數據，本文提供一種簡單的shell命令的方式批量刪除hbase里的數據。思路就是，建立hive與hbase的關聯表，通過hive sql查詢出符合條件的數據rowkey導出到本地文件，然后通過hbase shell批量刪除。示例如下:
1 創建hbase 表
#創建表名為“test:user_blacklist”,列簇名為“user”的hbase表
#hbase 表
create 'test:user_blacklist', 'user'
# phoneix映射表
create view "user_blacklist" ("rowkey" varchar primary key,"user"."dt" varchar,"user"."total_slave_count" varchar,"user"."api_slave_rate" varchar,"user"."active_7day_slave_count" varchar,"user"."active_7day_slave_weixin_rate" varchar,"device"."did" varchar,"device"."device_user_count" varchar,"device"."same_gyroscope_count" varchar);
2 創建hive映射表
CREATE EXTERNAL TABLE test.user_blacklist (
key string,
dt string,
total_slave_count string,
api_slave_rate string,
active_7day_slave_count string,
active_7day_slave_weixin_rate string,
device_user_count string,
same_gyroscope_count string
)
STORED BY 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
WITH SERDEPROPERTIES
("hbase.columns.mapping" =
":key,user:dt,user:total_slave_count,user:api_slave_rate,user:active_7day_slave_count,user:active_7day_slave_weixin_rate,device:device_user_count,device:same_gyroscope_count")
TBLPROPERTIES("hbase.table.name" = "test:user_blacklist");
3 通過hive sql導出指定rowkey刪除語句
hive -e "select concat('deleteall \'test:user_blacklist\',\'',key,'\'') from test.user_blacklist where dt>=20190708 and same_gyroscope_count is not null" > del_temp.txt
4 hbase shell批量刪除數據
hbase shell del_temp.txt > del.log