SQL Server Update with Inner Join

Question

I have 3 tables (simplified):

 tblOrder(OrderId INT)  
  tblVariety(VarietyId INT,Stock INT)  
  tblOrderItem(OrderId,VarietyId,Quantity INT)

If I place an order, I drop the stock level using this:

UPDATE tblVariety  
SET tblVariety.Stock = tblVariety.Stock - tblOrderItem.Quantity  
FROM tblVariety  
INNER JOIN tblOrderItem ON tblVariety.VarietyId = tblOrderItem.VarietyId  
INNER JOIN tblOrder ON tblOrderItem.OrderId = tblOrder.OrderId  
WHERE tblOrder.OrderId = 1

All fine, until there are two rows in tblOrderItem with the same VarietyId for the same OrderId. In this case, only one of the rows is used for the stock update. It seems to be doing a GROUP BY VarietyId in there somehow.

Can anyone shed some light? Many thanks.

Answer 1

You are doing an update. It will update once.

Edit: to solve, you can add in a subquery that will group your orderitems by orderid and varietyid, with a sum on the amount.

Answer 2

My guess is that because you have shown us simplified schema, some info is missing that would determine why have the repeated VarietyID values for a given OrderID.

When you have multiple rows, SQL Server will arbritrarily pick one of them for the update.

If this is the case, you need to group first

UPDATE V
SET
   Stock = Stock - foo.SumQuantity
FROM
    tblVariety V
    JOIN
    (SELECT SUM(Quantity) AS SumQuantity, VarietyID
     FROM tblOrderItem
      JOIN tblOrder ON tblOrderItem.OrderId = tblOrder.OrderId  
     WHERE tblOrder.OrderId = 1
     GROUP BY VarietyID
    ) foo ON V.VarietyId = foo.VarietyId  

If not, then the OrderItems table PK is wrong because if allows duplicate OrderID/VarietyID combinations (The PK should be OrderID/VarietyID, or these should be constrained unique)

Answer 3

From the documentation UPDATE

The results of an UPDATE statement are undefined if the statement includes a FROM clause that is not specified in such a way that only one value is available for each column occurrence that is updated (in other words, if the UPDATE statement is not deterministic). For example, given the UPDATE statement in the following script, both rows in table s meet the qualifications of the FROM clause in the UPDATE statement, but it is undefined which row from s is used to update the row in table t.

CREATE TABLE s (ColA INT, ColB DECIMAL(10,3))
GO
CREATE TABLE t (ColA INT PRIMARY KEY, ColB DECIMAL(10,3))
GO
INSERT INTO s VALUES(1, 10.0)
INSERT INTO s VALUES(1, 20.0)
INSERT INTO t VALUES(1, 0.0)
GO
UPDATE t 
SET t.ColB = t.ColB + s.ColB
FROM t INNER JOIN s ON (t.ColA = s.ColA)
GO