【发布时间】：2020-12-16 14:44:15
【问题描述】：
我有一个表“variables”和一个表“variables_history”，如下
create table variables
(
variables_id number,
variables_name varchar2(50),
variables_value varchar2(50),
variables_updated_at timestamp
);
create table variables_history
(
variables_id number,
variables_name varchar2(50),
variables_value varchar2(50),
variables_hist_updated_at timestamp
);
历史记录由触发器生成，如下所示。
CREATE OR REPLACE EDITIONABLE TRIGGER "myuser"."trigger_variables_update"
AFTER UPDATE ON myuser.variables
REFERENCING NEW AS NEW OLD AS OLD
FOR EACH ROW
BEGIN
IF :old.variables_value <> :new.variables_value THEN
INSERT INTO myuser.variables_history
(variables_id,
variables_name,
variables_value,
variables_hist_updated_at
VALUES
(variables_id,
:old.variables_name,
:old.variables_value,
old.variables_updated_at);
END IF;
END trigger_variables_update;
我还有一张桌子，上面有所有的维护保养
create table maintenance
(
maintenance_id number,
maintenance_status varchar2(20),
maintenance_date timestamp
);
我需要根据 maintenance_date 和 variables_updated_at 或 variables_hist_updated_at 的变量生成带有 maintenance_price 的输出
喜欢这个
WITH variables_data as
( SELECT 1 variables_id, 'maintenance_price' variables_name, '30.00' variables_value, '2020-08-01 05:00:00.000' variables_updated_at from dual),
variables_history_data as
(
SELECT 1 variables_id, 'maintenance_price' variables_name, '15.90' variables_value, '2019-10-01 11:30:00.000' variables_hist_updated_at from dual union all
SELECT 1 variables_id, 'maintenance_price' variables_name, '10.50' variables_value, '2020-01-01 01:00:00.000' variables_hist_updated_at from dual union all
SELECT 1 variables_id, 'maintenance_price' variables_name, '20.30' variables_value, '2020-05-01 12:30:00.000' variables_hist_updated_at from dual
),
maintenance_data as
(
SELECT 1 maintenance_id, 'COMPLETE' maintenance_status, '2019-02-01 00:30:00.000' maintenance_date from dual union all
SELECT 2 maintenance_id, 'COMPLETE' maintenance_status, '2019-05-01 01:30:00.000' maintenance_date from dual union all
SELECT 3 maintenance_id, 'COMPLETE' maintenance_status, '2019-11-01 02:30:00.000' maintenance_date from dual union all
SELECT 4 maintenance_id, 'COMPLETE' maintenance_status, '2020-07-10 05:30:00.000' maintenance_date from dual union all
SELECT 5 maintenance_id, 'FAILED' maintenance_status, '2020-08-02 11:30:00.000' maintenance_date from dual
SELECT 6 maintenance_id, 'COMPLETE' maintenance_status, '2020-08-20 11:30:00.000' maintenance_date from dual
)
Select
m.maintenance_id,
to_char(m.maintenance_date, 'yyyy/mm/dd') as maintenance_date
v.variables_value
from
maintenances m
join variables v on m.maintenance_date >= v.variables_updated_at
join variables_history vh on m.maintenance_date < variables_hist_updated_at
where maintenance_status = 'COMPLETE';
这个查询只是一个例子，我知道它错了
我需要这样的输出（并考虑变量可能有新的更新）。 “variable_value”需要是生成维护时的值。
maintenance_id | maintenance_date | variables_value |
---------------+------------------+-----------------+
1 | 2019-02-01| 15.90 |
---------------+------------------+-----------------+
2 | 2019-05-01| 15.90 |
---------------+------------------+-----------------+
3 | 2019-11-01| 10.50 |
---------------+------------------+-----------------+
4 | 2020-07-10| 20.30 |
---------------+------------------+-----------------+
6 | 2020-08-20| 30.00 |
---------------+------------------+-----------------+
【问题讨论】：
-
您的数据模型使这变得不必要地困难。您的历史记录表应包含该历史记录生效的时间戳。
-
为此，我需要在“variables_history”表（如“variables_created_at”）上创建一个新的时间戳列并更改触发器以保存 :old.variables_updated_at 值？
-
或者也许只是切换到触发器来保存 ":old.variables_updated_at" 值而不是设置当前的 "systimestamp"？
-
我正在考虑切换触发器以保存
:old.variables_updated_at而不是systimestamp。如果你想拥有“effective_from”（:old.variables_updated_at）和“effective_to”（systimestamp），拥有这两列并没有什么坏处。但你至少应该在历史表中拥有有效的起始日期才能做到这一点更容易。 -
谢谢。我会试试这样的。这是一个旧系统，处理起来有点复杂。
标签： sql oracle datetime lateral-join