Perfomance points in packages

Question

Hi all
sorry for asking an old question, just hoping to get some new answers, i recently went for an interview and i was asked "what are the advantage of packages?", i said
1.overloading,
2.encapsulation,
3.When you call a packaged subprogram for the first time, the whole package is loaded into memory. Therefore, subsequent calls to related subprograms in the package require no disk I/O
4.packages stop cascading dependencies and so avoid unnecessary recompiling
the interviewer said i'm missing one vital point in performace perspective and refused to tell that,
hi gurus can you kindly tell me what are the other points that can be added in the list of "advantages of packages in performance perspective"
Thank u

Answer 1

No. Procedures are basically used to create basic building blocks to extend the language - making it more flexible and powerful. And these are most definitely re-usable code.
Here is a basic example of how procedures are used in PL/SQL to extend the language by creating procedures to serve as "new language commands":
SQL>
SQL> -- a wrapper procedure that wraps raise_application_error
SQL> create or replace procedure RaiseError( errNo number, errMsg varchar2 ) is
2 begin
3 -- for debugging we can turn the following on: DBMS_OUTPUT.put_line( 'Error '||errNo||': '||errMsg );
4 -- we can also log the error to a log table via atonomous transaction, etc.
5 raise_application_error( -20000-errNo , errMsg );
6 end;
7 /
Procedure created.
SQL>
SQL> -- a utillity procedure to introduce a new command in PL/SQL
SQL> create or replace procedure Assert( condition boolean, errNo number, errMsg varchar2 ) is
2 begin
3 if not condition then
4 RaiseError( errNo, errMsg );
5 end if;
6 end;
7 /
Procedure created.
SQL>
SQL> create or replace type TMessageList is table of varchar2(200);
2 /
Type created.
SQL>
SQL>
SQL> -- now in a package, I can use these procedures/language extensions
SQL> create or replace package sampleP as
2
3 procedure DoSomething1( d date );
4
5 procedure DoSomething2( n number );
6 end;
7 /
Package created.
SQL>
SQL>
SQL> create or replace package body sampleP as
2
3 -- I define my error codes
4 ERR_DATE_IS_NULL constant number := 1;
5 ERR_DATE_INVALID constant number := 2;
6 ERR_EMP_TABLE_IS_EMPTY constant number := 3;
7
8 -- .. and a corresponding message for each error
9 ERROR_MESSAGE constant TMessageList := new TMessageList(
10 'The date may not be null.',
11 'The date may not be in the future.',
12 'The EMP table is empty.'
13 );
14
15
16
17 procedure DoSomething1( d date ) is
18 begin
19 -- the date may not be null
20 assert(
21 d is not null,
22 ERR_DATE_IS_NULL,
23 ERROR_MESSAGE( ERR_DATE_IS_NULL )
24 );
25
26 -- the date may not be in the future
27 assert(
28 d <= SYSDATE,
29 ERR_DATE_INVALID,
30 ERROR_MESSAGE( ERR_DATE_INVALID )
31 );
32
33 -- rest of the code for the proc...
34 end;
35
36 procedure DoSomething2( n number ) is
37 cnt integer;
38 begin
39 -- we do assertions to check our parameters, etc.
40 --
41 -- somewhere in our proc we select against EMP,
42 -- but the EMP table is empty and throws no data found
43 raise NO_DATA_FOUND;
44
45 -- more code...
46
47 exception when NO_DATA_FOUND then
48 -- our exception handler catches the exception and turns it
49 -- into a meaningful error
50 RaiseError( ERR_EMP_TABLE_IS_EMPTY, ERROR_MESSAGE(ERR_EMP_TABLE_IS_EMPTY) );
51 end;
52
53 begin
54 null;
55 end;
56 /
Package body created.
SQL>
SQL> var d varchar2(20)
SQL>
SQL> exec sampleP.DoSomething1( :d )
BEGIN sampleP.DoSomething1( :d ); END;
ERROR at line 1:
ORA-20001: The date may not be null.
SQL>
SQL> exec :d := SYSDATE+1;
PL/SQL procedure successfully completed.
SQL> exec sampleP.DoSomething1( :d )
BEGIN sampleP.DoSomething1( :d ); END;
ERROR at line 1:
ORA-20002: The date may not be in the future.
SQL>
SQL> exec sampleP.DoSomething2( 100 )
BEGIN sampleP.DoSomething2( 100 ); END;
ERROR at line 1:
ORA-20003: The EMP table is empty.
SQL>