Use result of upper query in union clause in lower query - oracle

Question

I am using union operator to combine results of 2 queries. Now I want to use the result of first query in second query so that I can exclude some records from second query.
E.g.
select <some_columns> from tableA Union select <same_columns> from tableA where <one_column_val> != <some_val_from_first_query>
Below is my query
SELECT a.*, b.*, c.*
FROM tableA a,
tableB b,
tableC c
where b.field_1 = c.field_1
and a.field_2 = c.field_2
union
SELECT a.*, b.*, c.*
FROM tableA a,
tableB b,
tableC c
where b.field_1 = c.field_1
and a.field_3=c.field_3
and a.field_2 <> {a.field_2 from upper query}
Please suggest the required alteration.
Thanks in advance

Answer 1

Each select in a SQL union has no knowledge of the others. So, you simply have to explicitly write out the entire query again, in your case either as a NOT EXISTS or a NOT IN. As an example, you can see the entire query copied as a derived table in a NOT EXISTS clause:
SELECT a.*, b.*, c.*
FROM tableA a,
tableB b,
tableC c
where b.field_1 = c.field_1
and a.field_2 = c.field_2
union
SELECT a.*, b.*, c.*
FROM tableA a,
tableB b,
tableC c
where b.field_1 = c.field_1
and a.field_3=c.field_3
and not exists (
select 1
from (
SELECT a.*, b.*, c.*
FROM tableA a,
tableB b,
tableC c
where b.field_1 = c.field_1
and a.field_2 = c.field_2
) x
where a.field_2 = x.field_2
)
If your RDBMS supports CTEs you can abstract the query into a CTE and not repeat yourself. Note you should probably also be using JOIN syntax if your RDBMS supports it, and you may have column name collisions with the * selector that you'll have to resolve.