I have a table1(DueDate varchar2(20)). It has thousands of date data in different format with some bad data like characters.
eg.
YYYYMMDD,
MM/DD/YYY
This is one of the reasons that storing date information in a character field is such a bad idea.
The easiest option is to create a function that attempts to convert the string to a date using the formats in whatever priority order you have (i.e. is 010203 Jan 2, 2003 or Feb 3, 2001 or something else) and catches the exceptions. Something like
CREATE OR REPLACE FUNCTION my_to_date( p_str IN VARCHAR2 )
RETURN DATE
IS
l_date DATE;
BEGIN
l_date := to_date( p_str, 'YYYYMMDD' );
RETURN l_date;
EXCEPTION
WHEN others THEN
BEGIN
l_date := to_date( p_str, 'MM/DD/YYYY' );
RETURN l_date;
EXCEPTION
WHEN others
THEN
RETURN null;
END;
END;
which works something like
SQL> CREATE OR REPLACE FUNCTION my_to_date( p_str IN VARCHAR2 )
2 RETURN DATE
3 IS
4 l_date DATE;
5 BEGIN
6 l_date := to_date( p_str, 'YYYYMMDD' );
7 RETURN l_date;
8 EXCEPTION
9 WHEN others THEN
10 BEGIN
11 l_date := to_date( p_str, 'MM/DD/YYYY' );
12 RETURN l_date;
13 EXCEPTION
14 WHEN others
15 THEN
16 RETURN null;
17 END;
18 END;
19 /
Function created.
SQL> select my_to_date( '19000101' ) from dual;
MY_TO_DAT
---------
01-JAN-00
SQL> ed
Wrote file afiedt.buf
1* select my_to_date( '01/02/2005' ) from dual
SQL> /
MY_TO_DAT
---------
02-JAN-05
Of course, you'd have to code the full set of valid date formats in your code, I'm just handling the first two in your list.