How to get a function to return a value occuring after a character

Question

I need to write a function to return the next value occurring after the ":" character in a column. If multiple values of ":" occurs then the function should return the sum of the next value occurring after each ":" in the column.
For example a rating value of 4:1 would return the value of 1. However, the rating of "5:1:1" should return a value of 1+1 = 2, and 6:2:1 will return of 2+1 = 3.
I have the below function skeletion and trying to figure out how the select statement will compute based on the position of : in the column and add the values and return the value back to function.
Function fn_check_internalrating(p_internalrating IN VARCHAR2)
      RETURN number
IS
    cnumber number;
    cursor c1 is
   select ................................
BEGIN
open c1;
fetch c1 into cnumber;
close c1;
RETURN cnumber;
EXCEPTION
WHEN OTHERS
THEN RETURN NULL;
END;

Answer 1

Hi,
You don't need a cursor: there's no table involved in this function, and no point in using any table.
Here's one way:
CREATE OR REPLACE FUNCTION  fn_check_internalrating
(   p_internalrating   IN   VARCHAR2
,   p_delimiter            IN   VARCHAR2     DEFAULT     ':'
RETURN  NUMBER
DETERMINISTIC          -- Same input always produces same output
IS
    cnumber     NUMBER     := 0;               -- value to be returned
    pos          PLS_INTEGER := INSTR ( p_internalrating
                                    , p_delimiter
                         );          -- position where delimiter was found
BEGIN
    WHILE  pos != 0
    LOOP
        cnumber := cnumber + TO_NUMBER ( SUBSTR ( p_internalrating
                                              , pos + 1
                                     , 1
     pos := INSTR ( p_internalrating
                      , p_delimiter
               , pos + 1
    END LOOP;
    RETURN cnumber;
END fn_check_internalrating;
SHOW ERRORSThis assumes the function is a stand-alone function. If it's part of a package, you don't say CREATE OR REPLACE at the beginning.
Try to make functions generic, so that if a similar (but not identical) situation comes up in 6 months from now, you can use the same function. I'm guessing that somethimes you may want to do the same thing with some character other than ':' before each number, so I added the 2nd (optional) argument p_delimiter. You can call the fucntion with either 1 or 2 arguments.
If an error occurs in a PL/SQL fucntion, an error message (showing the exact location of the error) is displayed, and execution halts. If you use an EXCEPTION sectinn, you lose all that functionality, or have to code it yourself. Only use an EXCEPTION handler when you really have to.
For this function, you may or may not want to. For example, if the character right after a delimiter is not a digit, the call to TO_NUMBER in function will raise "ORA-01722: invalid number". You may want to catch that error in an exception handler, and return 0 or NULL. On the other hand, you may want to test that the character after the delimiter is a digit before calling TO_NUMBER, and not have an EXCEPTION section.
What else could go wrong? Try to think of potential problems and fix them when you first write the function. If you discover an error next year, you'll have to spend a fair amount of time finding the function, and getting acquainted with it again.
What should the function return if p_internalrating is NULL, or doesn't contain any delimiters?
What if there's a number longer than 1 digit after a delimiter, e.g. '6:78:9'?