Can't find columns after declaration in SQL Server - sql

Question

I have a problem. I am creating a trigger which will prompt the user to not create a purchase order with a higher cost to their selling cost. I have declared a column on my script but its still showing this error:
Msg 207, Level 16, State 1, Procedure _trgZSCheckPrice, Line 31 [Batch Start Line 7]
Invalid column name 'fExclPrice'
The two columns are #Check and #Excl
I have attached my code below any help is advisable:
SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
CREATE TRIGGER _trgZSCheckPrice
ON [dbo].[_btblInvoiceLines]
FOR INSERT, UPDATE
AS
BEGIN
SET NOCOUNT ON;
DECLARE
#Valid int,
#ValidPO int,
#DocType int,
#DocState int,
#Check int,
#Excl float,
#POPrice float
SELECT
#Check = ubIICheck,
#Excl = fExclPrice,
#POPrice = fUnitPriceExcl
FROM
INSERTED
SELECT
#Excl = fExclPrice,
#Check = ubIICheck
FROM
stkitem A
INNER JOIN
_etblPriceListPrices B ON A.StockLink = B.iStockID
SELECT
#POPrice = fUnitPriceExcl
FROM
_btblInvoiceLines C
LEFT JOIN
InvNum D ON C.iInvoiceID = D.AutoIndex
BEGIN
IF (#DocType = 5 AND #DocState <> 7) AND #CHECK = 1
BEGIN
IF #Excl > #POPrice
BEGIN
RAISERROR ('Message from Management:
You are not allowed to Purchase above Selling Cost.
The transaction will be rolled back. ', 16, 1)
ROLLBACK TRANSACTION
END
END
END
END

Answer 1

It seems like ideally, here, you'd be implementing this as a multi-table CHECK constraint (in fact, in standard SQL, these are called ASSERTIONs. So far as I'm aware, only Postgre implements them).
If you're happy to trade away the ability to completely control the error message, I'd usually prefer to implement this in a declarative manner rather than relying on a trigger.
Here's how to do such a check with an indexed view. First a couple of tables that are somewhat like your problem domain (but not very fleshed out since you didn't put any definitions in your question):
create table dbo.ListPrices (
ID int not null,
Price decimal(12,4) not null,
constraint PK_ListPrices PRIMARY KEY (ID)
)
create table dbo.Orders (
ID int not null,
ListPriceID int not null,
MyPrice decimal(12,4) not null,
constraint PK_Orders PRIMARY KEY (ID),
constraint FK_Orders_ListPrices FOREIGN KEY (ListPriceID)
references dbo.ListPrices (ID)
)
go
insert into dbo.ListPrices (ID,Price) values (1,12.50),(2,25.00)
And now we create a special helper table. It's not needed if you already have a suitable table (such as a numbers table) in your database:
create table dbo.Two (
n int not null,
constraint PK_Two PRIMARY KEY (n),
constraint CK_Two_Only CHECK (n in (1,2))
)
go
insert into dbo.Two (n) values (1),(2)
And now we create the view:
go
create view dbo.DRI_NoOrderPricesOverListPrices
with schemabinding
as
select
1 as p /* Constant */
from
dbo.Two t
cross join
dbo.Orders o
inner join
dbo.ListPrices lp
on
o.ListPriceID = lp.ID
where
/*Conditions for failure*/
o.MyPrice > lp.Price
go
create unique clustered index IX_NoOrderPricesOverListPrices
on DRI_NoOrderPricesOverListPrices(p)
The key feature here is that the view joins all of the tables we're interested in together and we can put whatever conditions we like in the where clause, referencing multiple tables, comparing column values, etc.
The conditions we're specifying is for what shouldn't be allowed in the database. So, we're saying we should be able to insert a row in Orders provided that its price is less than or equal to the price in line items.
And we can. This insert succeeds:
insert into Orders (ID,ListPriceID,MyPrice) values (1,2,17.00)
And this one fails:
insert into Orders (ID,ListPriceID,MyPrice) values (2,1,17.00)
Msg 2601, Level 14, State 1, Line 42
Cannot insert duplicate key row in object 'dbo.DRI_NoOrderPricesOverListPrices' with unique index 'IX_NoOrderPricesOverListPrices'. The duplicate key value is (1).
(As I said at the top, we don't get so much control over the error message - but making a good choice of name here should make it reasonable to deduce what's happening and directly exposing SQL Server error messages to the users is something I'd generally try to avoid)