I\'m working with an old SQL 2000 database and I don\'t have a whole lot of SQL experience under my belt. When a new row is added to one of my tables I need to assign a defa
So, for example, in a TAG table (where tags are applied to posts) if you want to count one tag as another...but default to counting new tags as themselves, you would have a trigger like this:
CREATE TRIGGER [dbo].[TR_Tag_Insert]
ON [dbo].[Tag]
AFTER INSERT
AS
BEGIN
SET NOCOUNT ON;
UPDATE dbo.Tag
SET [CountAs] = I.[ID]
FROM INSERTED AS I
WHERE I.[CountAs] IS NULL
AND dbo.Tag.ID = I.ID
END
Use a trigger as suggested by Stephen Wrighton:
CREATE TRIGGER [myTable_TriggerName] ON dbo.myTable FOR INSERT
AS
SET NOCOUNT ON
UPDATE myTable
SET
timeValue = '2 hours' -- assuming string values
where ID in (
select ID
from INSERTED
where
timeValue = ''
AND workCategory = 'A'
)
Yeah, trigger.
Naturally, instead of hard-coding the defaults, you'll look them up from a table.
Expanding on this, your new table then becomes the work_category table (id, name, default_hours), and you original table maintains a foreign key to it, transforming fom (id, work_category, hours) to (id, work_category_id, hours).
Generally I steer away from triggers. Almost all dbms have some sort of support for constraints.
I find them easier to understand , debug and maintain.
If what you are looking for is to define a column definition based on another column you can do something like this:
create table testable
(
c1 int,
c2 datetime default getdate(),
c3 as year(c2)
);
insert into testable (c1) select 1
select * from testable;
Your result set should look like this :
c1 | c2 | c3
1 | 2013-04-03 17:18:43.897 | 2013
As you can see AS (in the column definition) does the trick ;) Hope it helped.
I would use a trigger on Insert.
Just check to see if a value has been assigned, and if not, go grab the correct one and use it.