About Optimization Of PL/SQL Constructs

Question

Hi All,
I am asking very silly question, some one asked to me how to optimize pl/sql procedure or function.i had explained all the things for optimization of the sql query or we can put the pl/sql subprogram into package.
But from the face expression of the person,i think he wants more from my side,but really i did not know any thing except of these things.
I want to know ideas of the experts on this issue.
Thanks a lot

Answer 1

If by optimization that person was asking about techniques then there are a few general rules. Keeping in mind that general rules are starting points and require validation is every specific case.
1. Never use PL/SQL to do what you can do with SQL.
2. Never use single-row processing when you can process a set
So, for example, you can write three anonymous blocks to do some piece of work ... one that uses SQL such as:
BEGIN
  INSERT INTO t1
  SELECT * FRM t2;
  COMMIT;
END;
Then another that does the same thing with BULK COLLECT and FORALL and finally a third using a cursor loop. Which will be fastest and more efficient doing the same work is reasonably certain.
The second interpretation of the question would be how does one determine the efficiency of PL/SQL and for that I would refer you to the built-in DBMS_PROFILER (10g and before) and DBMS_HPROF (11g) packages.
There are a number of authors that have written very good materials on the subject among which I would suggest you explore asktom.oracle.com, Tom Kyte's many excellent books and Steven Feuerstein. There are many others but this will give you a good starting point.