Eirikur Eiriksson (11/18/2014)

---

serg-52 (11/18/2014)

---

kriskumark96 (11/17/2014)

---

Hi,

The three queries is of the same tables with different where condition are used , so that we cannot have a join or cross join on it

Or

Is there any way to add 2 different queries as a single resultant

select col1 from tb1 where col2=''

select col4 from tb1 where col3=''

the resultant should be col1,col4

Just a guess, if the first select returns

A

B

C

, and the second one returns

1

2

do you need

A 1

B 2

C NULL

?

Then number the results as you need and full-join them by those numberings.

select t1.col1, t2.col4

from (select row_number() over (order by ...) as n, col1

from tb1 where col2='') as t1

full outer join (select row_number() over (order by ...) as n, col4

from tb1 where col3='') as t2 on t1.n=t2.n

you need to choose how the results should be ordered.

Quick thought, this will indeed produce a set which has the desired columns but those values will be paired/matched by the over/order clause in the individual queries, hence the cardinality and the value distribution will control the "implied" relationship. If the over/order clause columns are the same in all query instances, those columns are then effectively the common key for the set and can therefore be used as such!

😎