Using a HAVING clause in an UPDATE statement

问题

This query

SELECT
FirstName, LastName, NCAAStats.AccountId, College_Translator.school_name, StatTypeId, COUNT(*) AS 'Count'
FROM NCAAstats
INNER JOIN College_Translator
ON College_Translator.AccountID = NCAAstats.AccountId
GROUP BY FirstName, LastName, NCAAStats.AccountId, College_Translator.school_name, CalendarYear, StatTypeId
HAVING COUNT(*) >1
ORDER BY 'Count' DESC

Selects records that I would like to set an ISValid bit to 0.

These records are records that appear twice in my database due to an input error.

I'm looking for something like:

UPDATE NCAAstats
SET IsValid = 0
WHERE (my select statement)

This is on MS SQL SERVER 2008

Thanks!

回答1:

You can join to that subquery like so:

update n1 set
    isvalid = 0
from
    ncaastats n1
    inner join (
        SELECT
        FirstName, LastName, NCAAStats.AccountId, College_Translator.school_name, StatTypeId, COUNT(*) AS 'Count'
        FROM NCAAstats
        INNER JOIN College_Translator
        ON College_Translator.AccountID = NCAAstats.AccountId
        GROUP BY FirstName, LastName, NCAAStats.AccountId, College_Translator.school_name, CalendarYear, StatTypeId
        HAVING COUNT(*) >1
    ) n2 on
        n1.accountid = n2.accountid

回答2:

SQL Server can do updates like:

UPDATE table SET col=vaue
FROM (
  SELECT ......
)

You should look here first:

http://msdn.microsoft.com/en-us/library/aa260662(v=sql.80).aspx

回答3:

The above are good suggestions.... here's another easy way to do it :

update ncaastats set isvalid = 0
where accountId in (
    SELECT AccountId
    FROM NCAAstats
    INNER JOIN College_Translator
    ON College_Translator.AccountID = NCAAstats.AccountId
    GROUP BY FirstName, LastName, NCAAStats.AccountId, College_Translator.school_name, CalendarYear, StatTypeId
    HAVING COUNT(*) >1
)

** Forgive me if I messed up the columns name, but you get the idea.

回答4:

Use a CTE, and do what is basically a self join

;with NCAAstatsToUpdate(
    SELECT AccountId 
    FROM NCAAstats n
        INNER JOIN College_Translator ct
      ON ct.AccountID = n.AccountId 
    GROUP BY FirstName, LastName, n.AccountId, ct.school_name, 
         CalendarYear, StatTypeId 
    HAVING COUNT(*) >1 )
UPDATE NCAAstats 
SET IsValid=0
FROM NCAAstats n
inner join NCAAstatsToUpdate u
    on n.AccountId = u.AccountId

Or better yet, use the windowing functions.

;with NCStats as(
 Select distinct row_number() over (partition by FirstName, LastName, n.AccountId, ct.school_name, 
         CalendarYear, StatTypeId order by n.accountId) rw, n.*
 FROM NCAAstats n
        INNER JOIN College_Translator ct
      ON ct.AccountID = n.AccountId 
)
Update NCStats
Set IsValid=0
Where rw>1

Note that second does not update the "first" record to invalid, and that it assumes that there that there is a 1 to 1 relationship between NCAAstats and College_Translator.

回答5:

For SQL Server 17

UPDATE table SET col = val 
(SELECT cols FROM table .. )

来源：https://stackoverflow.com/questions/8793914/using-a-having-clause-in-an-update-statement

标签

sql

sql-server-2008

having