蜂巢：失败：在多个表/子查询中发现了SemanticException列

编程问答 2022-08-12

问题描述

我正在尝试按以下方式运行查询

hive -e "set hive.execution.engine=mr;set hive.strict.checks.cartesian.product=false;
set hive.mapred.mode=nonstrict;use db1; select col1,col2 from tb1 where col_date='2020-08-15' and col3='Y' 
and col4='val4' and col1 not in 
( select distinct col1 from db2.tb2 where col_date='2020-08-15' and 
col5='val5' and col6='val6' and col3='Y' and col4='val4') "

但我不断得到

Failed: SemanticException Column col1 Found in more than One Tables/Subqueries

我在做什么错？我该如何解决？

db1.tb1中的列

col1
col2
col_date
col3
col4

db2.tb2中的列

col1
col2
col_date
col3
col4
col5
col6

解决方法

为表添加别名，并将其用于所有列：

hive -e "set hive.execution.engine=mr;set hive.strict.checks.cartesian.product=false;
set hive.mapred.mode=nonstrict;
use db1; 
select t1.col1,t1.col2 
  from tb1 t1
 where t1.col_date='2020-08-15' and t1.col3='Y' and t1.col4='val4' 
   and t1.col1 not in 
( select distinct t2.col1 from db2.tb2 t2 
   where t2.col_date='2020-08-15' and t2.col5='val5' and t2.col6='val6' and t2.col3='Y' and t2.col4='val4' ) "

或者，如果您的Hive版本不支持NOT IN子查询，则可以对同一对象使用LEFT JOIN +过滤器

hive -e "set hive.execution.engine=mr;set hive.strict.checks.cartesian.product=false;
set hive.mapred.mode=nonstrict;
use db1; 
select t1.col1,t1.col2 
  from tb1 t1
       left join 
        ( select distinct t2.col1 from db2.tb2 t2 
           where t2.col_date='2020-08-15' 
             and t2.col5='val5' 
             and t2.col6='val6' 
             and t2.col3='Y' 
             and t2.col4='val4' 
        ) s on t1.col1 = s.col1 
 where t1.col_date='2020-08-15' and t1.col3='Y' and t1.col4='val4' 
   and s.col1 is null --filter out joined records
"
hivehiveqlsqlsql