select g.dt,z.integ from (select * from @dtz where integ=0)z right join @dt g on g.dt=z.dtz

talking bout joins is there a way to join these two result sets :

TABLE 1

DATE
01/01/2008
02/01/2008
03/01/2008

TABLE 2
DATE EMP
01/01/2008 123
02/01/2008 123
03/01/2008 456

with condition of Emp=’123′

into these
RESULTS:

DATE EMP
01/01/2008 123
02/01/2008 123
03/01/2008 NULL

thanks…

I was expecting this article to explore the effect of the order of join statements in a query and how the order might be best used to optimise it.

Here is what I think:
I believe it is the way the join operators organize the rows when joining the two tables together. In the “outer join” query the execution plan shows hash match operators, if you change the hash match operator to nested loop on EmployeeDepartmentHistory and Department you get the same ordered result set as the “inner join” query.

I don’t know enough about how hash match join to comment on why it yeilds a different order than nested set so maybe someone can help me with that.

Here is the modified “outer join” query that yeilds the same ordered result set as the “inner join” query.

SELECT he.EmployeeID, he.Title, hd.Name, hd.GroupName, hdh.StartDate
FROM HumanResources.Employee he
LEFT JOIN HumanResources.EmployeeDepartmentHistory hdh
ON he.EmployeeID = hdh.EmployeeID
RIGHT JOIN HumanResources.Department hd
ON hd.DepartmentID = hdh.DepartmentID OPTION (LOOP JOIN)

If an ORDER BY statement is not used, then the order is undefined. You might be able to predict it, but to base code/logic on it is a bad, bad thing to do.