Question

let’s say I’ve got two queries:

select top 20 idField, field1, field2 from table1
where idField in
    (
    select idField from table1 where field3 like '...'
    union
    select idField from table2 where field4 like '...'
    )
order by sortfield1

select top 20 idField, field1, field2 from table1
where field3 like '...' 
or idfield in 
(select idField from table2 where field4 like '...')
order by sortfield1

let’s say that both tables are rather large (table2 larger than table1),
average range of results number is 0-400, but I always want max 20 results.

Is it possible to say with big probability which statement would be more performant
and/or what would the choice of statement depend on?
Thank you

Answer 1

In the perfect world this should not make a difference, but I would still try out this approach too. Could be faster:

select top 20 idField, field1, field2 from table1
where idField in
    (
    select top 20 idField from table1 where field3 like '...'
    union
    select top 20 idField from table2 where field4 like '...'
    )