Select name without special characters

Question

I have a table with names, some of which contain special characters('/','\','#','&'). I need to select the names columns with certain conditions.
1. The name with these special characters should not be picked, esp. when another same name (without the special characters exists
e.g: if I have 2 entries ABC& and ABC, I should select ABC
2. If no entries are clear of this special characters, pick the last name.
e.g: if I have 2 entries ABC& and ABC#, I should select ABC#.
Please help me device the code for this.

Answer 1

Hi,
Try something like this ...
select version from v$instance ;
<br>
VERSION
10.1.0.3.0
<br>
1 row selected.
<br>
create table test_values ( id number, val varchar2(10) );
<br>
insert into test_values values ( 1, 'abc' );
insert into test_values values ( 2, 'abc' );
insert into test_values values ( 3, 'abc/' );
insert into test_values values ( 4, 'abc\' );
insert into test_values values ( 5, 'abc#' );
insert into test_values values ( 6, 'def' );
insert into test_values values ( 7, 'def' );
insert into test_values values ( 8, '#ghi#' );
insert into test_values values ( 9, '/ghi/' );
insert into test_values values ( 10, null );
<br>
select id, val from test_values ;
<br>
        ID VAL
         1 abc
         2 abc
         3 abc/
         4 abc\
         5 abc#
         6 def
         7 def
         8 #ghi#
         9 /ghi/
        10
<br>
10 rows selected.
<br>
select
  max(id) keep
    ( dense_rank last
      order by instr(standardized_val,'#') desc, id
    ) as target_id,
  max(val) keep
    ( dense_rank last
      order by instr(standardized_val,'#') desc, id
    ) as target_val
from
  ( select id, val,
      translate( val, '/\#&', '####' ) standardized_val,
      translate( val, 'x/\#&', 'x' ) clean_val
    from test_values
  ) tv
group by
  clean_val
<br>
TARGET_ID TARGET_VAL
         2 abc
         7 def
         9 /ghi/
        10
<br>
4 rows selected.Message was edited by: jfuda - added tags for formatting source