Oracle 通过一个表字段,更新另一个表字段:
方法一:通过update实现
update g_sn_status A
set A.out_process_time = (select B.rec_time from g_sn_defect B where B.serial_number = A.serial_number and B.rp_status is null)
where A.serial_number in
(
select distinct A.serial_number from g_sn_status A
inner join g_sn_defect B on A.serial_number = B.serial_number
where A.out_process_time <> B.rec_time and B.rp_status is null and A.process_id = B.process_id and A.current_status = 1
)
如果没有where约束部分,A表的out_process_time字段将会全部更新。【A表存在、B表不存在serial_number时,A表的out_process_time字段会更新成null】
方法二:通过merge into实现
merge into g_sn_status A using g_sn_defect B on (A.serial_number = B.serial_number) when matched then update set A.out_process_time = B.rec_time where A.serial_number like '2018%' when not matched then insert (A.serial_number,A.out_process_time) values (B.serial_number,B.rec_time) where B.rec_time between to_date('2020051008','YYYYMMDDHH24') and to_date('2020052020','YYYYMMDDHH24');
作用:判断B表和A表是否满足ON中条件,如果满足则用B表去更新A表,如果不满足,则将B表数据插入A表但是有很多可选项,如下:
1.正常模式
2.只update或者只insert
3.带条件的update或带条件的insert
4.带delete的update【鸡肋,而且会增加SQL语言的复杂度】
该SQL类似SQLPLUS的case when...语句,想想也知道为啥有很多可选项了。
总结:方法二虽然功能更强大,但是用的少。