SQL Trigger insert problems - sql

Question

I want to make a trigger where I can check if the value from the stock of a product 0 is.
Then the trigger should make the value 1.
My Trigger
CREATE TRIGGER [IfStockIsNull]
ON [dbo].[Products]
FOR DELETE, INSERT, UPDATE
AS
BEGIN
if [Products].Stock = 0
UPDATE [Products] SET [Stock] = 1
FROM [Products] AS m
INNER JOIN inserted AS i
ON m.ProductId = i.ProductId;
ELSE
raiserror('Niks aan het handje',10,16);
END
I'm getting the error:
Altering [dbo].[IfStockIsNull]...
(53,1): SQL72014: .Net SqlClient Data Provider: Msg 4104, Level 16, State 1,
Procedure IfStockIsNull, Line 7 The multi-part identifier "Products.Stock"
could not be bound.
(47,0): SQL72045: Script execution error. The executed script:
ALTER TRIGGER [IfStockIsNull]
ON [dbo].[Products]
FOR DELETE, INSERT, UPDATE
AS BEGIN
IF [Products].Stock = 0
UPDATE [Products]
SET [Stock] = 1
FROM [Products] AS m
INNER JOIN
inserted AS i
ON m.ProductId = i.ProductId;
ELSE
RAISERROR ('Niks aan het handje', 10, 16);
END
An error occurred while the batch was being executed.
Maybe you guys can help me out?

Answer 1

A number of issues and surprises in what you're trying to run here. But basically, don't try to run procedural steps when you're dealing with sets. inserted can contain 0, 1 or multiple rows, so which row's stock are you asking about in your IF?
Better to deal with it in the WHERE clause:
CREATE TRIGGER [IfStockIsNull]
ON [dbo].[Products]
FOR INSERT, UPDATE --Removed DELETE, because ?!?
AS
BEGIN
UPDATE [Products] SET [Stock] = 1
FROM [Products] AS m
INNER JOIN inserted AS i
ON m.ProductId = i.ProductId;
WHERE m.Stock = 0
--Not sure what the error is for - the above update may have updated
--some number of rows, between 0 and the number in inserted.
--What circumstances should produce an error then?
END
Simple demo script that the UPDATE is correctly targetting only matching rows from inserted:
declare #t table (ID int not null, Val int not null)
insert into #t(ID,Val) values (1,1),(2,2),(3,3)
update
#t
set
Val = 4
from
#t t
inner join
(select 2 as c) n
on
t.ID = n.c
select * from #t
Shows that only the row with ID of 2 is updated.
Even Microsoft's own example of UPDATE ... FROM syntax uses the UPDATE <table> ... FROM <table> <alias> ... syntax that Gordon asserts doesn't work:
USE AdventureWorks2012;
GO
UPDATE Sales.SalesPerson
SET SalesYTD = SalesYTD + SubTotal
FROM Sales.SalesPerson AS sp
JOIN Sales.SalesOrderHeader AS so
ON sp.BusinessEntityID = so.SalesPersonID
AND so.OrderDate = (SELECT MAX(OrderDate)
FROM Sales.SalesOrderHeader
WHERE SalesPersonID = sp.BusinessEntityID);
GO
This example does have an issue which is further explained below it but that relates to the "multiple matching rows in other tables will result in only a single update" flaw that anyone working with UPDATE ... FROM should make themselves aware of.

Answer 2

If you want to stop the insert if any rows are bad, then do that before making the changes:
ALTER TRIGGER [IfStockIsNull] ON [dbo].[Products]
FOR INSERT, UPDATE -- DELETE is not really appropriate
AS BEGIN
IF (EXISTS (SELECT 1
FROM Products p JOIN
inserted i
ON m.ProductId = i.ProductId
WHERE p.Stock = 0
)
)
BEGIN
RAISERROR ('Niks aan het handje', 10, 16);
END;
UPDATE p
SET Stock = 1
FROM Products p INNER JOIN
inserted AS i
ON p.ProductId = i.ProductId;
END;
Notes:
The comparison in the IF checks if any of the Product rows have a 0. If so, the entire transaction is rolled back.
You might complain that you cannot reject a single row. But that is how transactions work. If any row fails in an INSERT, the entire INSERT is rolled back, not just a single row.
The UPDATE is not correct, because you have Products in the FROM clause and the UPDATE clause. You need to use the alias for the first one.