Help with sorting a query

Question

Hi,
I have one table (ctry) with a two columns as below
ctry - varchar2(2)
zone - varchar2(6)
SE
SE1:1
SE
SE1:2
SE
SE1:3
SE
SE1:4
SE
SE1:10
SE
SE1:11
SE
SE1:5
If I sort as below
select ctry, zone
from ctry
order by ctry, zone
I got result as below
ctry - varchar2(2)
zone - varchar2(6)
SE
SE1:1
SE
SE1:10
SE
SE1:11
SE
SE1:2
SE
SE1:3
SE
SE1:4
SE
SE1:5
I want result as below
ctry - varchar2(2)
zone - varchar2(6)
SE
SE1:1
SE
SE1:2
SE
SE1:3
SE
SE1:4
SE
SE1:5
SE
SE1:10
SE
SE1:11
Is that possible at all when type is varchar2?
Example table/data:
create table ctry
ctry varchar2(2),
zone varchar2(6)
insert into ctry values('SE', 'SE1:1');
insert into ctry values('SE', 'SE1:2');
insert into ctry values('SE', 'SE1:3');
insert into ctry values('SE', 'SE1:4');
insert into ctry values('SE', 'SE1:10');
insert into ctry values('SE', 'SE1:11');
insert into ctry values('SE', 'SE1:5');

Answer 1

Hi,
So, you have a number stored at the end of a VARCHAR2 column, and you want to threat it as a number when sorting.
Pleiadian and Blushadow showed hos to actually convert that part of the string to a NUMBER.  If the part of the string after the ':' is not a number, even if that happens on only 1 row, row will get no results other than an error message.  To guard against that, you can pad just the part after the ':' with '0's, like this:
SELECT    *
FROM      ctry
ORDER BY  REGEXP_REPLACE ( REGEXP_REPLACE ( zone
                                          , '(\d+$)'
                                          , '00000\1'
                         , '0+(\d{6})'
                         , '\1'
The inner REGEXP_REPLACE adds 5 '0's before the group of digits at the end of the string.  (Since the column is defined as VARCHAR2 (6), you'll never have to add more than 5 '0's to get all the sub-strings to the same length.)
The outer REGEXP_REPLACE removes extra '0's, so that all the numbers that were padded to more than 6 digits are now exactly 6 digits.
As Blushadow siad, this is only a way of coping with the problem, not solving it.  If the part of the zone after the ':' needs to be treated like a NUMBER, then it should be in a NUMBER column.  Cramming it into the same column with 'SE1' seems to violate First Normal Form.  Why not split the zone into 2, or possibly more, columns?  If you must leave it as 1 column, can you store it in a useable form, so that you don;t need to ask questions like this?  If you must store the data exactly as posted, then add another column that reflect the sort order.