Reputation: 57
This is SQL Server 2016. I have the following data in only one table:
custID | prodID | title | titleCount | isMasterTitle
--------+--------+--------+-------------+-----------
266 | 191750 | prod01 | 1 | 0
266 | 191750 | prod02 | 4 | 0
266 | 191750 | prod03 | 25 | 0
300 | 20125 | prod04 | 3 | 0
300 | 20125 | prod05 | 15 | 0
I want to group by custID, prodID and title and update isMasterTitle field to 1 for every max() titleCount per group.
So, I want the following:
custID | prodID | title | titleCount | isMasterTitle
--------+--------+----------+------------+---------------
266 | 191750 | prod01 | 1 | 0
266 | 191750 | prod02 | 4 | 0
266 | 191750 | prod03 | 25 | 1
300 | 20125 | prod04 | 3 | 0
300 | 20125 | prod05 | 15 | 1
I'm trying the following:
UPDATE [dbo].[_Variations]
SET isMasterTitle = 1
FROM [dbo].[_Variations] v1
INNER JOIN (SELECT custID, prodID, MAX(titleCount) AS mtitleCount
FROM [_Variations]
GROUP BY custID,prodID) as v2 ON v1.custID = v2.custID and v1.prodID = v2.prodID and v1.titleCount = v2.mtitleCount
Upvotes: 1
Views: 156
Reputation: 3744
try the following:
;with cte
as
(
select isMasterTitle, ROW_NUMBER() over (partition by custID, prodID order by titleCount desc) rn
from @t
)
update cte
set isMasterTitle = 1
where rn = 1
select * from @t
Your given code also works fine.
Please find the db<>fiddle here.
Upvotes: 1
Reputation: 222462
I would recommend leveraging a powerful feature of SQL Server called the updateable common-table-expression.
You can build a cte that uses window functions to identify which row should be updated, and then directly update it; there is no need to join again the original table in the outer query. This makes the query both shorter and more efficient:
with cte as (
select
isMaster,
row_number() over(partition by custID, prodID order by titleCount desc) rn
from [dbo].[_Variations]
)
update cte set isMaster = 1 where rn = 1
Upvotes: 1