Question about correct use of WM_CONCAT

Question

Good morning,
I'm trying to accomplish the following:
starting with this
     ENAME
     ADAMS
     ALLEN
     BLAKE
...and ending with this
     OLD_NAME   NEW_NAME
     ADAMS      AADMS
     ALLEN      AELLN
     BLAKE      ABEKL
...basically, alphabetically sorting the characters inside each name. I have an intermediate step that looks promising
select e.ename oldname, substr(e.ename, iter.pos, 1) newcharpos
  from emp e,
       (select rownum pos from emp) iter
where iter.pos <= length(e.ename)
order by e.ename, substr(e.ename, iter.pos, 1);The above yields
OLDNAME    NEWCHARPOS
ADAMS      A
ADAMS      A
ADAMS      D
ADAMS      M
ADAMS      S
ALLEN      A
ALLEN      E
ALLEN      L
ALLEN      L
ALLEN      N
BLAKE      A
BLAKE      B
BLAKE      E
BLAKE      K
BLAKE      L
...the individual characters are in the right sequence, I figured all I had to do was to use WM_CONCAT (and replace the comma it inserts with the empty string) and I'd be done. I figured this would do it: (replace of comma left out for clarity)
select oldname,
       wm_concat(newcharpos) newname
  from (
        select e.ename oldname, substr(e.ename, iter.pos, 1) newcharpos
          from emp e,
               (select rownum pos from emp) iter
         where iter.pos <= length(e.ename)
         order by e.ename, substr(e.ename, iter.pos, 1)
  group by oldname;but the sequence of newcharpos gets messed up in the process and, instead of the expected result, I get this:
OLDNAME    NEWNAME
ADAMS      A,S,M,D,A
ALLEN      A,N,L,L,E
BLAKE      A,L,K,E,B
...My question is, how do I structure the last query so the character order stays as calculated in the inner query ?
Thank you for your help,
John.

Answer 1

Or
SQL> select   oldname,
       wm_concat(newcharpos) keep (dense_rank last order by null)  newname
  from (
        select e.ename oldname, substr(e.ename, iter.pos, 1) newcharpos
          from emp e,
               (select rownum pos from emp) iter
         where iter.pos <= length(e.ename)
         order by e.ename, substr(e.ename, iter.pos, 1)
  group by oldname
OLDNAME    NEWNAME                      
ADAMS      A,A,D,M,S                    
ALLEN      A,E,L,L,N                    
BLAKE      A,B,E,K,L                    
CLARK      A,C,K,L,R                    
FORD       D,F,O,R                      
JAMES      A,E,J,M,S                    
JONES      E,J,N,O,S                    
KING       G,I,K,N                      
MARTIN     A,I,M,N,R,T                  
MILLER     E,I,L,L,M,R                  
SCOTT      C,O,S,T,T                    
SMITH      H,I,M,S,T                    
TURNER     E,N,R,R,T,U                  
WARD       A,D,R,W                      
14 rows selected.Or (11gR2)
SQL> select ename oldname, column_value newname
  from emp,
       xmltable(('string-join(for $i in (' || rtrim(regexp_replace(ename, '(.)', '"\1",'),',') || ') order by $i return $i, "")'))
OLDNAME    NEWNAME                      
SMITH      HIMST                        
ALLEN      AELLN                        
WARD       ADRW                         
JONES      EJNOS                        
MARTIN     AIMNRT                       
BLAKE      ABEKL                        
CLARK      ACKLR                        
SCOTT      COSTT                        
KING       GIKN                         
TURNER     ENRRTU                       
ADAMS      AADMS                        
JAMES      AEJMS                        
FORD       DFOR                         
MILLER     EILLMR                       
14 rows selected.