Question

2 related questions :

---

question #1 :

I have a myTable with id , name , lastname

id is seed (pk).

name has index on it. //not nullAble.

which query is faster ?

• select id , name , lastname from mytable where name = name //(name has index on it!)

or

• select id , name , lastname from mytable

---

question #2 : *now lets assume name is nullable.*

from the first sample

there is a select query sp which its parameter is @name + @lastName

(returning all rows where name = @name or lastName= @lastName)

which query is faster ?

select id , name , lastname from mytable where 
   (@name is null or  name=@name)

and

   (@lastName is null or  lastName=@lastName) 

here if @name or @lastName are null – so it wont compare name=@name

or

 select id , name , lastname from mytable where 
       (name =  isNull(@name,name))

    and

      (lastName=  isNull(@lastName,lastName))

here , if @name or @lastName are null , so there STILL will be equalization
of name= name and lastName=lastName , which i think is bad
performance.

please help ?

Answer 1

As written, the first two queries are equivalent.

name = name will always evaluate true, and will be a constant evaluation basically. Both of those queries will run a table scan.

The second two queries should be equivalent as well.

ISNULL may be slightly faster but I doubt it – the optimizer expands it out to a CASE statement so it will likely short circuit.