操作顺序SQL Server中的内部联接和Where子句性能？

Question

Is there any Performance problem to use operation in miss order?

Like 1. All Inner join first then all where condition later.

    select * from
    t1
    inner join t2 on t1.t2Id = t2.Id
    inner join t3 on t1.t3Id = t3.Id
    inner join t4 on t2.t4Id = t4.Id
    where 
    t1.Id in (1,2,3,4,5)
    and t2.Id in (1,2,3,4,5,6,7)
    and t3.Name like '%a' 

2. All table with Respective Where and then Inner join

 select * from
    (select * from t1 where t1.Id in (1,2,3,4,5)) a
    inner join (select * from t2 where t2.Id in (1,2,3,4,5,6,7)) a1 on a.t2Id = 
    a.Id
    inner join (select * from t3 where t3.Name like '%a') a2 on a.t3Id = a2.Id
    inner join t4 on a1.t4Id = t4.Id

It may effect on query Performance?

3. Also Order of Where Condition?

   Like

    select * from t1 inner join t2 on t1.t2Id = t2.Id where t1.t2Id in (1,2,3,4,5,6) and t2.t3Id in (1,2,3,4,5) 

Answer 1

A SQL query goes through three phases when it is run:

1. The query is parsed (and the various references are looked up).
2. The execution plan is created, with an optimization phase based on what the query needs to accomplish.
3. The query plan is execution.

As a result of the optimization, the way you write the query often has less effect on the performance than you might think. Lots of people have worked very hard on figuring out the best way to optimize queries -- and there are probably lots of things that you are not even aware of (such as different join algorithms, join ordering, pushing down expression evaluations, and so on).

For your examples, the SQL Server optimizer should produce the same execution plans. The engine is smart enough to realize that these are really doing the same thing.

Note: This is not true of all query engines. Some have pretty poor optimizers, and there would be differences in performance.