Output parameter in a mapping

Question

I want to check a table if there are records in that table and return a indc as 'Y' or 'N' depending on the records selected and return that indc as the putput parameter in a mapping. Can this be done in a mapping or I have no other choice but to create a database function and call the funtion in the process flow.
thanks

Answer 1

Well - still using the store proceduire - one thing you can do is to check in your pre-map procedure and raise an exception if you do not want to run the mapping. The downside would be that the mapping would return a failure status which would be problematic for process flows (if you are using process flows to automate your project). If you are not using process flows for automation you could implement the logic in your custom handler to check the returned exception code and if it were your code then continue to the next mapping.
Alternately you could build your logic as a function to be included in your automation (e.g. embedded as a sql object in a process flow) which then branches to either run the mapping or not depending on the returned value.
Or, if you don't like branching, you could simply build the check into a wrapper function which then calls the mapping as part of it's functionality. This wrapper is then what you would call in your automation.
e.g.
procedure check_and_run is
   p_status varchar2(100);
   dummy   number;
begin
   select distinct 1 into dummy
   from your_test_table
   where your_condition = your_test_value;
   -- we will get here if a row exists, otherwise the select will have failed.
   your_mapping_name.main(p_status);
exception
    when no_data_found then
        -- select failed. Do not run mapping.
        dbms_output.put_line('Skipped the mapping as it failed the go/no-go test');
end;
/ I suppose you could create this sort of thing as a sql script to run in a process flow instead of as a stored item - if you are dead set against creating any objects (although why that would be I don't know. After all, OWB is going to be creating objects as it is.)
You have options. Which one you choose is up to you based on your architecture and preferences.
But I'm really not sure why you are so concerned with creating code objects. The entire OWB concept revolves around the notion of creating objects. Why put arbitrary bounds on it?
Cheers,
Mike