Question

I have the following trigger:

BEGIN
DECLARE @email varchar(200)
DECLARE @jobcode int
DECLARE @status char(1)

DECLARE @emaild varchar(200)
DECLARE @jobcoded int
DECLARE @statusd char(1)

SET @statusd = NULL

SELECT @status = z.status, @email = p.EMail, @jobcode = z.jobID
    FROM zipoutfiles z
    INNER JOIN inserted AS i ON z.jobID = i.jobID
    INNER JOIN PS_LoginUser AS p ON z.UserID = p.UserID

SELECT @statusd = z.status, @emaild = p.EMail, @jobcoded = z.jobID
    FROM zipoutfiles z
    INNER JOIN deleted AS d ON z.jobID = d.jobID
    INNER JOIN PS_LoginUser AS p ON z.UserID = p.UserID
    WHERE d.jobID = @jobcode


IF ((@status = 'D' AND @statusd = 'R') OR (@status = 'D' AND @statusd = 'E'))
BEGIN
    EXEC SendMail @email, @jobcode
END
END

I want to be able to run SendMail when status goes from E to D or R to D, but not D to D (if it gets updated again) and also when it gets inserted as D. What am I doing wrong here:

Answer 1

Not sure what your table schemas are, but this may get you all of the appropriate emails:

select p.EMail as Email, z.JobId as JobCode
  from ZipOutFiles as ZOF inner join
    inserted as i on i.JobId = ZOF.JobId inner join
    PS_LoginUser as PLU on PLU.UserId = ZOF.UserId left outer join
    deleted as d on d.JobId = ZOF.JobId
    where
      ( d.Status = 'E' and i.Status = 'D' ) or -- E -> D.
      ( d.Status = 'R' and i.Status = 'D' ) or -- R -> D.
      ( d.Status is NULL and i.Status = 'D' ) -- Inserted D.

I’ve assumed that you aren’t really updating JobId. If so, how do you match the before and after rows?

Also assumed is that Status cannot be NULL. If so, the last condition needs to be modified to properly detect no corresponding row was found in the deleted table.