Ah! I see. when you raise an exception, you basically “go to” the corresponding exception block.

APC: Before I started in Oracle I worked on a particular flavour of COBOL that didn’t support conditional ERFORMs. GOTOs were a way of life.

I started my programming career with COBOL, that was in 1992. COBOL and GOTO were best buddies.

In PL/SQL, I just cannot see a real need to use labels and GOTO’s. Labels can be easily replaced with comments. GOTO’s can be easily replaced with many other programming controls that provide a much better clarity to the program logic.

SQL> Declare
  2    LN$I pls_integer := 0 ;
  3    LN$J pls_integer := 0 ;
  4  Begin
  5    <<boucle1>>
  6    Loop
  7      LN$I := LN$I + 1 ;
  8      Loop
  9         LN$J := LN$J + 1 ;
 10         dbms_output.put_line( to_char( LN$I ) || ',' || to_char( LN$J ) ) ;
 11         EXIT boucle1 WHEN LN$J > 3 ;
 12      End loop ;
 13    End loop ;
 14  End ;
 15  /
1,1
1,2
1,3
1,4

Procédure PL/SQL terminée avec succès.

I don’t think he will ever be able to revoke it given that he passed away few years back

They have? Nobody tells me anything.

Actually I was thinking about the NTK newsletter and not some computer science pronouncement. I don’t think Dijkstra has revoked his fatwa. Still, if you’ve ever had to follow the possible paths through a function that has over a dozen RETURN statements you might find yourself pining for a nice GOTO exit_point.

GOTOs are like triggers: there’s nothing inherently wrong with the construct, it’s simply the potential for abuse that renders their use suspect. Simply avoiding the use of GOTO statement does not make our code well-structured.

Of course not. I was merely pointing out that raising an exception is a special case of GOTO. Particularly given a case like this:

    ....
    do_something_a;
    do_something_b;
EXCEPTION
   WHEN others THEN blah;

If do_something_a raises an exception to be handle by the calling program then essentially there is an invisible GOTO that bypasses do_something_b.

Before I started in Oracle I worked on a particular flavour of COBOL that didn’t support conditional ERFORMs. GOTOs were a way of life. So I know bad use of GOTO create spaghetti code. I also know that well structured GOTOs have their place.

Of course, Oracle PL/SQL has a panoply of control structures and I have never had to use a GOTO since I stopped COBOLing. Indeed the other day I found myself with a LOOP that had more than one EXIT statement and felt unclean.

I take it you use GOTO to raise exceptions

They have? Nobody tells me anything.

Cheers

Tim…