背景需求:用一条sql查询没有任何关联的两张表,并且查询中涉及到了聚合函数,查询结果需要作为一条数据(列转行)。
两张没有关联的表放到一个结果集中,此处考虑使用union all,
select count(1) bcrzc,0 bsjgj,opt_time from t_manual_entry_workload a where opt_type = 'CIQ入暂存' and is_delete = 0
union all
select 0 bcrzc,count(1) bsjgj,op_date opt_time from til_inspection_rec where brand = '保时捷' and step = '港检' and is_deleted = 1
结果如下:
简单的union all查出来是以不同数据行的形式显示,需要列转行,外边就加了一层查询
select sum(a.bcrzc) bcrzc, sum(a.bsjgj) bsjgj from(
select count(1) bcrzc,0 bsjgj,opt_time from t_manual_entry_workload a where opt_type = 'CIQ入暂存' and is_delete = 0
union all
select 0 bcrzc,count(1) bsjgj,op_date opt_time from til_inspection_rec where brand = '保时捷' and step = '港检' and is_deleted = 1) a
where a.opt_time >= '2017-01-01' and a.opt_time <= '2017-11-30'
第二种查询方法:
select (select count(1) from t_manual_entry_workload a where opt_type = 'CIQ入暂存' and opt_time >= '2017-01-01' and opt_time <= '2017-11-30') AAA
,(select count(1) from til_inspection_rec where brand = '保时捷' and step = '港检' and op_date >= '2017-01-01' and op_date <= '2017-11-30') BBB
虽然第二种查询看起来稍微简单容易理解一些,但是很明显效率没有第一种高,猜想是由于where后的时间条件重复校验导致的。
版权声明:本文为yangfengjueqi原创文章,遵循 CC 4.0 BY-SA 版权协议,转载请附上原文出处链接和本声明。