Question

I need help to optmize the SQL logic in one of my functions. Please, note that I am not able to use store procedure.

Here is my table. It will be initialized using @MainTable that contains a lot of records.

DECLARE TABLE @ResultTable
(
    ResultValue INT
)

These are tables that stores some parameters – they can be emty too.

DECLARE TABLE @ParameterOne (ParameterOne INT)
DECLARE TABLE @ParameterTwo (ParameterOne NVARCHAR(100))
...
DECLARE TABLE @ParameterN(ParameterN TINYINT)

Now, I need to join a lot of tables to my @MainTable in order to select from it only some of its records.
The selected records depend on the information stored in the parameters table.

So, my current solution is:

INSERT INTO  ResultTable(ResultValue)
SELECT ResultValue 
FROM MainTable M
INNER JOIN @MainOne MO
    ON M.ID=MO.ID
....
INNER JOIN @MainN MN
    ON M.IDN=MN.ID
WHERE (EXISTS (SELECT 1 FROM @ParameterOne WHERE ParameterOne=MO.ID) OR NOT EXISTS (SELECT 1 FROM @ParameterOne))
       AND 
      ...
       AND
      (EXISTS (SELECT 1 FROM @ParameterN WHERE ParameterN=MN.Name) OR NOT EXISTS (SELECT 1 FROM @ParameterN ))

So, the idea is to add the records only if they match the current criteria from the parameters tables.

Because I am not able to use procedure to build dynamic query I am using the WHERE clause with combinations of EXISTS and NOT EXISTS for each parameter table.

The problem is that it works slower when I am adding more and more parameters table. Is there an other way to do this without using a lot of IF/ELSE statements checking what parameter table has records – it will make the function a lot bigger and difficult for read.

And ideas and advices are welcomed.

Answer 1

Good question.

Try the following one:

INSERT INTO  ResultTable(ResultValue)
SELECT ResultValue 
FROM MainTable M
INNER JOIN (SELECT * FROM @MainOne WHERE (EXISTS (SELECT 1 FROM @ParameterOne WHERE ParameterOne=@MainOne.ID) OR NOT EXISTS (SELECT 1 FROM @ParameterOne))) MO
    ON M.ID=MO.ID
....
INNER JOIN (SELECT * FROM @MainN WHERE (EXISTS (SELECT 1 FROM @ParameterN WHERE ParameterOne=@MainN.Name OR NOT EXISTS (SELECT 1 FROM @ParameterN))) MO
    ON M.IDN=MN.ID

Advantages:

1. Result of the JOIN is more quickly, because it does not process all data (it is already filtered)
2. It looks more simple for adjusting