Stored procedure vs anonymouse block

Question

Hi all,
I had an interesting problem and don't know why.
My procedure takes the files from a directory and process them. using vx_xmltype.toobject(vo_object,..) I load xml files into sql object.
create type in_obj as object (types varchar2(100));
create type in_obj_nt as varray(1000) as in_obj;
create type final as object (name varchar2(100),objects in_obj_nt);
And my code was something like this:
for i in ( all file_names) loop
   vx_xml := ....;
   declare
      obj final;
   begin
      xml.toobject(vo_obj,...); 
      ---process obj--
   end;
end loop;Starting with second file in the loop the in_obj from vo_obj was initialized with null values(not the ones from the file).
But when I changed the code with this one:
for i in ( all file_names) loop
   vx_xml := ....;
   xml_to_obj(vx_xml,obj);
/*   declare
      obj final;
   begin
      xml.toobject(obj,...);
   end;*/
      ---process obj--
end loop;I replaced the declare..begin..end block with a stored procedure that does the same thing...the objects were correctly initialized.
What is the difference between these two examples?
there is a difference in how the memory is assigned for objects? can you provide me some documentation in this direction?
thanks
Message was edited by:
alinux

Answer 1

I think it is a scope problem, although you haven't included enough code to be certain...
you are declaring obj within the anonymous block. It will be out of scope after the termination of that block, which is when you try to process it.
When you changed to a function you would have had to declare obj earlier, and it would still be in scope when you got to the processing part.