T-SQL递归更新

Question

当我更新帐户的收入或结果值时，我应该使用什么查询来更新总计和 Total_previous 列。请注意，记录按 Posting_DataTime 列排序，Transaction_ID 未排序。

Account No	Transaction_ID	Posting_DataTime	Income	Outcome	Total	Total_previous
001	11111111	2021-12-01 00:00	100	null	100	null
001	11112331	2021-12-04 04:00	10	null	110	100
021	11111031	2021-12-04 04:23	1200	null	2000	800
001	11111231	2021-12-08 07:44	null	40	70	110
001	00011231	2021-12-20 11:59	200	null	270	70

如果我将 11112331 (Transaction_ID) 的收入值从 10 更新为 50，则该帐户 (001) 的下一行应该如下所示

Account No	Transaction_ID	Posting_DataTime	Income	Outcome	Total	Total_previous
001	11111111	2021-12-01 00:00	100	null	100	null
001	11112331	2021-12-04 04:00	50	null	150	100
021	11111031	2021-12-04 04:23	1200	null	2000	800
001	11111231	2021-12-08 07:44	null	40	110	150
001	00011231	2021-12-20 11:59	200	null	310	110

Answer 1

您似乎希望将 posting_datatime 由 account_no 排序的（收入 - 结果）相加。一个更好的选择而不是更新是将其作为“视图”而不是每次有变化时都更新。

    create table t(account_no varchar(10),trans_id varchar(1000),posting_datatime datetime,Income int, Outcome int);

    insert
      into t
    select '001','11111111','2021-12-01 00:00',100  ,null union all
    select '001','11112331','2021-12-04 04:00',10   ,null union all
    select '021','11111031','2021-12-04 04:23',1200 ,null union all
    select '001','11111231','2021-12-08 07:44',null ,40  union all
    select '001','00011231','2021-12-20 11:59',200  ,null;

    select account_no
           ,trans_id
           ,posting_datatime
           ,income
           ,outcome
           ,sum(isnull(income,0)-isnull(outcome,0)) 
               over(partition by account_no order by posting_datatime) as total
           ,sum(isnull(income,0)-isnull(outcome,0)) 
               over(partition by account_no order by posting_datatime)
                 -(isnull(income,0)-isnull(outcome,0))as total_prev
      from t
    order by posting_datatime;

account_no	trans_id	posting_datatime	income	outcome	total	total_prev
001	11111111	2021-12-01 00:00:00.000	100	(null)	100	0
001	11112331	2021-12-04 04:00:00.000	10	(null)	110	100
021	11111031	2021-12-04 04:23:00.000	1200	(null)	1200	0
001	11111231	2021-12-08 07:44:00.000	(null)	40	70	110
001	00011231	2021-12-20 11:59:00.000	200	(null)	270	70

link https://sqlize.online/sql/mssql2017/e86a41327cd213441797c775652f9a69/

T-SQL递归更新

T-SQL recursive update

tsql

sql-server