hive HQL 练习 from....insert into

insert into table us_mac_stat partition(day='20170619') select mobile_user_id, c from (select mobile_user_id, count(distinct mac) c from default.mobile_request_log_orc where day_id='20170619' and mobile_user_id<>'' and mobile_user_id is not null and mac<>'' and mac is not null group by (mobile_user_id)  t where t.c>1;


insert into table us_ip_stat partition(day='20170621',h) select mobile_user_id, c, hour_id from ( select mobile_user_id, count(distinct ip) c , hour_id from default.mobile_request_log_orc where day_id='20170619' and mobile_user_id<>'' and mobile_user_id is not null and ip<>'' and ip is not null group by mobile_user_id,hour_id) t where t.c>2; 


from ...
inster into.. 
inster into ..



合并
from ( select mobile_user_id,ip,mac,hour_id from default.mobile_request_log_orc where  day_id='20170619' and mobile_user_id<>'' and mobile_user_id is not null) publicf
 insert into table us_mac_stat partition(day='20170619') select  mobile_user_id, count(distinct mac) c  where  mac<>'' and mac is not null group by  mobile_user_id having count(distinct mac)>1
 insert into table us_ip_stat partition(day='20170619',h)  select  mobile_user_id, count(distinct ip) cip , hour_id where  ip<>'' and ip is not  null group by mobile_user_id,hour_id having count(distinct ip)>2;


执行后，发现h分区里一个个文件都很小，分件数量很多，需要手动设置一下参数

 set mapred.reduce.tasks = 50; 

根据自己的数据量大小调整