Question

我在hive中有两个巨大的表，我有一个运行这种类型的查询：示例查询：

    select
    employee_id,
    employee_name,
    employee_address,
    employee_join_date
    employee_travel_pincodes
    from employee
    where 
    employee_join_date = (select join-date from hr_records)
    and
    employee_travel_pincodes in (select _pincodes from hr_records) //returns     multiple records

在hive中实现此目的的最佳方法是什么，我可以使用子查询，但看起来它不是实现所需输出的最干净的方法我正在使用hive 0.13

Answer 1

我建议使用JOIN以获得更好的性能：

SELECT 
    employee_id,
    employee_name,
    employee_address,
    employee_join_date
    employee_travel_pincodes
FROM employee e
INNER JOIN hr_records hr
    ON e.employee_join_date = hr.join-date
    AND e.employee_travel_pincodes = hr._pincodes

Answer 2

您可以使用左半连接，可以在Hive中使用，代替：

select * from table1 where columnx in (select column x from table2);

这与左半连接相同：

 select columnx from table1 a left outer join table2 b on a.columnx=b.columnx;

因此我会根据您所需查询的描述使用：

select
  employee_id,
  employee_name,
  employee_address,
  employee_join_date
  employee_travel_pincodes
from employee a
left semi join hr_records b on a.employee_travel_pincodes=b._pincodes
where a.employee_join_date=b.join-date
;

如何使用列值等于选择其他查询来运行配置单元查询

2 个答案: