Convert varchar2 field into date formatted: DD-MON-YYYY

Question

Thanks in advance for anyone's help on this matter as I know it takes your time and expertise. I am pretty new to SQL but learning my way through it just have an issue with a text to date field conversion. It is an Oracle 10g database and I am writing in SQL. There is a field called Demand which is formatted in varchar2 format of ddmmyy. There is also a field that is formatted as a date called Payment which is formatted as DD-MON-YYYY.
Essentially I need to do a simple Payment >= Demand, however as you can see that is some issue with that being a varchar2 field. Does anyone know if it is possible to do that type of expression against those two fields. Was thinking about possibly converting the varchar2 to a date but not sure how to get to that DD-MON-YYYY format.
Also there are situations where this Demand field will often times be null as it would have never recieved any outbound correspondence in the past and would not have a date at all.
Thanks
Edited by: user10860766 on Aug 18, 2009 8:14 AM
Edited by: user10860766 on Aug 18, 2009 8:19 AM

Answer 1

Hi,
It's hard to detect bad dates in pure SQL, especially if you need to be precise about when February 29 is valid.
It's easy with a user-define function, like the one in [this thread|http://forums.oracle.com/forums/thread.jspa?messageID=3669932&#3669932].
Edited by: Frank Kulash on Aug 18, 2009 3:50 PM
To create a stand-alone function:
CREATE OR REPLACE FUNCTION     to_dt
(     in_txt          IN     VARCHAR2                    -- to be converted
,     in_fmt_txt     IN     VARCHAR2     DEFAULT     'DD-MON-YYYY'     -- optional format
,     in_err_dt     IN     DATE          DEFAULT     NULL
RETURN DATE
DETERMINISTIC
AS
BEGIN
     -- Try to convert in_txt to a DATE.  If it works, fine.
     RETURN     TO_DATE (in_txt, in_fmt_txt);
EXCEPTION     -- If TO_DATE caused an error, then this is not a valid DATE: return in_err_dt
     WHEN OTHERS
     THEN
          RETURN in_err_dt;
END     to_dt
/To use it:
SELECT  primary_key  -- and/or other columns to identify the row
,       demand
FROM    table_x
WHERE   demand          IS NOT NULL
AND     to_dt ( demand
           , 'DDMMYY'
           )          IS NULL;