How to run a stored procedure in TOAD 9.0

Question

Dear Friends,
I am using TOAD 9.0 .
Running a procedure but getting the error ..
create or replace PROCEDURE UPD_TBL(I_table_name IN VARCHAR2) IS
CURSOR c_columns is select column_name from ALL_TAB_COLUMNS where
TABLE_NAME = I_table_name;
BEGIN
FOR rec in c_columns
LOOP
IF ( rec.COLUMN_NAME = 'X_NNM_TYPE_CD' ) THEN
UPDATE I_table_name set X_NNM_TYPE_CD = REPLACE (X_NNM_TYPE_CD, '-', '_')
WHERE INSTR (X_NNM_TYPE_CD, '-') > 0 ;
ELSIF ( rec.COLUMN_NAME = 'X_REVENUE_TYPE_CD' ) THEN
UPDATE I_table_name set X_REVENUE_TYPE_CD = REPLACE (X_REVENUE_TYPE_CD, '-', '_')
WHERE INSTR (X_REVENUE_TYPE_CD, '-') > 0 ;
END IF;
END LOOP;
EXCEPTION
when OTHERS then
NULL;
END UPD_TBL;
ERROR: TABLE OR VIEW DOES NOT EXIST.
Please let me know how to run the procedure in toad.
thanks/kumar

Answer 1

Kumar,
At a minimum, you would need to remove both the exception handler in your procedure and the exception handler in your anonymous block. As has been mentioned before, the code
EXCEPTION
  WHEN OTHERS THEN
    NULL;
END;is almost always a bug. In your case, your code is almost certainly generating an exception and your exception handler is silently catching and ignoring that exception. If you care whether or not any rows are updated, that is a bug.
Once you remove the exception handlers, you'll very likely discover that the SQL statements you are generating and executing dynamically are malformed. Off the top of my head, it would appear that there is no space between the table name and the SET keyword in the string you are constructing. That means that if I_TABLE_NAME is 'EMP', for example, you'd construct a statement like
UPDATE empSET x_nnm_typ_cd = ...'Oracle would interpret empSET as the table name, it would interpret X_NNM_TYP_CD as a table alias, and then it would encounter the equals sign unexpectedly and generate a syntax error. You'd need to insert a space after the table name and before the SET keyword (note where folks inserted extra spaces in their sample code).
In general, you are probably better served by having a local variable that you use to build up the SQL statement. That way, you can print (or log) the SQL statement that is constructed. When you encounter a syntax error at runtime, that will generally make debugging easier. So you'd probably want something like
  l_sql_stmt VARCHAR2(1000);
BEGIN
  FOR rec IN cur_c_columns LOOP
    IF (rec.column_name = 'X_NNM_TYPE_CD') THEN
      l_sql_stmt :=
         'UPDATE '||i_table_name||
         ' SET x_nnm_type_cd = REPLACE(x_nnm_type_cd, ''-'', ''_'')'||
         'WHERE INSTR(x_nnm_type_cd, ''-'') > 0';
      dbms_output.put_line( l_sql_stmt );
      EXECUTE IMMEDIATE l_sql_stmt;
...When you get a syntax error, you can then take a look at the SQL statement that was generated and debug the problem more easily.
Justin