Oracle 9i function to perform data manipulation

Question

Hi everyone,
I have two scenarios. I would like to write one Oracle 9i function that can accommodate both these scenarios, using PL/SQL developer. I intend passing a varchar2 parameter to this function. Within the function, I perform a select on a table and see if this value sent has a parent record.
If it does not have a parent record (null or blank) then I take this value, join with another table, perform some processing and return one varchar2 value.
If it has a parent record (not null or not blank) then I need to retrieve all the other children records associated with this same parent record, take this value, join with another table,perform some processing and return one varchar2 value.
I have two simple queries that I wrote: one for record with null/blank parent, and the other for record with same parent and other children. These two queries work individually, when I combine them is where I run into problems. Sample table:
TABLE XYZ data: (note: if there is a parent, parent shows up in both parent and child field in the table)
child parent
1234
5107 3167
6259 3652
5678 9754
3167 3167
3652 3652
1649
2306 9754
4657 3652
9754 9754
8753 3652
passing record 1234 into the function.
SELECT A.child, A.parent, C.field1
FROM OTHERTABLE C,
(SELECT TableA.child, TableA.parent
FROM XYZ TableA
WHERE TableA.child ='1234')A
WHERE A.child = C.field2
----(output of this query is 1 record with 3 fields 1234 and null/blank and 987654321. I place this output into a cursor and then take 987654321 and place into a variable and return this value)
passing record 4657 into the function
SELECT c.CHILDfield, c.PARENTfield, TableOther.field1
FROM OTHERTABLE TableOther,(
(SELECT B.CHILDfield, B.PARENTfield FROM
(SELECT A.PARENTfield, i.CHILDfield FROM TableB,
(SELECT TableA.CHILDfield, TableA.PARENTfield
FROM XYZ TableA
WHERE TableA.CHILDfield ('4657' ))A
-- pass 4657 and then get the parent
WHERE A.PARENTfield = TableB.PARENTfield)B
-- gets associated children with the parent from previous query.
-- note: parent can exist in the parent and child fields
WHERE B.CHILDfield <> B.PARENTfield))C
--strips off the parent and gets child/children only as I don't need
--the parent when it is also a child
WHERE C.CHILDfield = TableOther.field1
--(output of this query is 3 records
4657 3652 912345671
6259 3652 945362798
8753 3652 934274857)
I place this output into a cursor and then have a CASE statement that goes thru each record, and appends the third value into a variable, the ultimate return value of this will be as follows:
912345671;945362798;934274857)
The dilemma is writing a generic query to accommodate both these situations. My current function can handle each of these situations individually, but not both. For testing purposes, I change the select statement depending on whether I have a record that has a parent on not!
Any help is greatly appreciated.

Answer 1

You want one function to handle two cases? This is the best opportunity to use the IF STATEMENT.
FUNCTION MyFunc (p_input IN VARCHAR2) RETURN VARCHAR2 IS
ReturnValue VARCHAR2;
BEGIN
IF <blank or null test> THEN
SELECT ... -- Blank or null select statement
... -- Blank or null data processing
RETURN ReturnValue;
ELSE
SELECT ... -- Other select statement
... -- Other data processing
RETURN ReturnValue;
END IF;
END MyFunc;