Need inner and outer join in one function - oracle

Question

I have one table that holds a record for each customer (main table). I then have a table with additional detail for some customers. The additional detail table sometimes has no records for a record in the main table. Sometimes the detail table has multiple records for a record in the main table & if this is the case I need the most recent record (hence the max subselect).
The trouble is my function only returns values for the few records in the detail table. If I comment out the portion of the function that looks at the detail table and just return the STAT3 value it seems to work. How do I make the second select statment below only apply if there is a result for that query?
create or replace FUNCTION "F_RETURN_STAT" (
N_UNIQUE IN NUMBER)
RETURN VARCHAR2
IS
V_STAT3 varchar2(20);
V_STAT varchar2(20);
V_STAT2 varchar2(20);
D_ACTDATE date;
D_STARTDATE date;
BEGIN
select expire into D_ACTDATE
from main_table a
where a.uniquefield = N_UNIQUE;
IF
D_ACTDATE > SYSDATE
or
D_ACTDATE is null
then
V_STAT :='TRUE';
else
v_STAT :='FALSE';
end if;
select b.startdate into D_STARTDATE
from main_table a, detail_table b
where a.uniquefield= b.main_table_id(+) and
b.main_table_id = N_UNIQUE and
b.uniquefield in
(select max(c.uniquefield) from detail_table c group by main_table_id);
if
D_STARTDATE is not null
then
V_STAT2 :='FALSE';
end if;
if
V_STAT2 ='FALSE'
then
V_STAT3 :='FALSE';
ELSE
V_STAT3 := V_STAT;
end if ;
RETURN(V_STAT3);
end;

Answer 1

I think this version of your function will solve your problem:
CREATE OR REPLACE FUNCTION f_return_stat(n_unique IN NUMBER)
RETURN VARCHAR2 IS
v_stat3 VARCHAR2(20);
v_stat VARCHAR2(20);
v_stat2 VARCHAR2(20);
d_actdate DATE;
d_startdate DATE;
BEGIN
--First Query
SELECT expire
INTO d_actdate
FROM main_table a
WHERE a.uniquefield = n_unique;
IF d_actdate > SYSDATE OR d_actdate IS NULL THEN
v_stat := 'TRUE';
ELSE
v_stat := 'FALSE';
END IF;
BEGIN
--Second Query
SELECT b.startdate
INTO d_startdate
FROM detail_table b
WHERE b.main_table_id = n_unique
AND b.uniquefield IN (SELECT MAX(c.uniquefield)
FROM detail_table c
GROUP BY main_table_id);
EXCEPTION
WHEN NO_DATA_FOUND THEN
d_startdate := NULL;
END;
IF d_startdate IS NOT NULL THEN
v_stat2 := 'FALSE';
END IF;
IF v_stat2 = 'FALSE' THEN
v_stat3 := 'FALSE';
ELSE
v_stat3 := v_stat;
END IF;
RETURN (v_stat3);
END;
In your version of the second query, your join (a.uniquefield= b.main_table_id) and your filter (b.main_table_id = N_UNIQUE) are equivalent, so main_table a can be removed altogether. The only reason to leave it in is to make sure that your query always returns a row. If you use exception handling to catch the NO_DATA_FOUND exception, that need goes away and you can simplify your query to just select from detail_table b.

Answer 2

I believe there could be a more efficient way however this might do the job:
SELECT b.startdate
INTO d_startdate
FROM detail_table b
WHERE b.main_table_id = n_unique
and
(b.uniquefield in
(select max(c.uniquefield) from detail_table c group by main_table_id)
or b.uniquefield is null);