I have a table in SQL Server that has a PK (ID
) and another (logical) primary key made by a couple of other columns (although there is no UNIQUE constraint on that). Let's say, table PERSON
, PK = PERSON_ID
, then NAME
, SURNAME
, AGE
I'd like it was possible to say
UPDATE PERSON SET AGE = 43 WHERE NAME = 'XX' AND SURNAME = 'YYY'
and have it executed only if 'updated rows' = 1, otherwise (more than 1 row) NO EXECUTION at all. The problem is that I'm not sure if NAME and SURNAME uniquely identify a record, and I have no way to tell it a priori.
Ideas?
Try the below query... it will help you
UPDATE PERSON
SET AGE = 43
WHERE NAME = 'XX'
AND SURNAME = 'YYY'
AND 1 = (SELECT COUNT(*) FROM PERSON WHERE NAME = 'XX' AND SURNAME = 'YYY)
I'd put the conditional before the update rather than in the where clause.
IF (SELECT COUNT(*) FROM PERSON WHERE NAME = 'XX' AND SURNAME = 'YYY') = 1
UPDATE PERSON SET AGE = 43 WHERE NAME = 'XX' AND SURNAME = 'YYY'
Rather than writing a complex WHERE clause or IF statement, I usually just wrap the whole thing in a transaction and check @@ROWCOUNT:
BEGIN TRAN
UPDATE PERSON SET AGE = 43 WHERE NAME = 'XX' AND SURNAME = 'YYY'
IF @@ROWCOUNT > 1 ROLLBACK TRAN ELSE COMMIT TRAN
something like that should do it.
UPDATE Person p1
SET p1.AGE = 43
WHERE NAME = 'XX' AND SURNAME = 'YY'
AND NOT EXISTS (
SELECT NULL FROM Person p2
WHERE p1.NAME = p2.NAME
AND p1.SURNAME = p2.SURNAME
GROUP BY p2.NAME, p2.SURNAME
HAVING COUNT(*) > 1)
UPDATE Person p1
SET p1.AGE = 43
WHERE NAME = 'XX' AND SURNAME = 'YY'
AND NOT EXISTS (
SELECT NULL FROM Person p2
WHERE p1.NAME = p2.NAME
AND p1.SURNAME = p2.SURNAME
AND p1.ID <> p2.ID)