Help with Stored Procs.

Question

Hello All,
Orcl Version:
Oracle Database 11g Enterprise Edition Release 11.2.0.2.0 - 64bit Production
PL/SQL Release 11.2.0.2.0 - Production
"CORE     11.2.0.2.0     Production"
TNS for Linux: Version 11.2.0.2.0 - Production
NLSRTL Version 11.2.0.2.0 - Production
I have a scenario, where I've to run the procedures from a package multiple times with different input parameter(key) primary key of one of the tables used inside the proc.
For ex,
pkg1
proc1(Key);
proc2(Key);
proc3(Key);
proc30(Key);
Engine(Main code) runs these procs multiple times with different params. Sometimes a proc runs more than 4000 times with different param value(key). I have 2 questions.
1. A proc when run for the first time, it is compiled and for the next time it will be in compiled form. I'm not sure how it works for the second run. Is the query in the proc cached and helps improve the performance?
2. Are there any ways to improve the performance?
Like, I'm hitting the base tables everytime a proc is run. Is there any advantage of creating MV's or something like that?
Thanks
Shank.

Answer 1

Hi, Shank,
SamFisher wrote:
Orcl Version:
Oracle Database 11g Enterprise Edition Release 11.2.0.2.0 - 64bit Production ...
Thanks for including that!
I have a scenario, where I've to run the procedures from a package multiple times with different input parameter(key) primary key of one of the tables used inside the proc.
For ex,
pkg1
proc1(Key);
proc2(Key);
proc3(Key);
proc30(Key);
Engine(Main code) runs these procs multiple times with different params. Sometimes a proc runs more than 4000 times with different param value(key). I have 2 questions.
1. A proc when run for the first time, it is compiled and for the next time it will be in compiled form. Not usually. A stored procedure is compiled when it is created, that is, when you run the CREATE OR REPLACE PROCEDURE (or CREATE OR REPLACE PACKAGE) command.
It may be automatically re-compiled when you call it the first time. For example, if procedure A calls some other procedure B, and B has been re-compiled since A was created, then A will be re-compiled when you run it the first time.
I'm not sure how it works for the second run. Is the query in the proc cached and helps improve the performance? Performance may be a little faster after the first call, because the procedure, and the objects it references (such as table blocks) may already be cached in memory.
2. Are there any ways to improve the performance?
Like, I'm hitting the base tables everytime a proc is run. Is there any advantage of creating MV's or something like that? If the package body contains a BEGIN section, then that section is run the first time the package is used , but not again in the same session. You can use that section to initialize package variables (or SYS_CONTEXT variables, or global temporary tables) so you don't have to re-compute them every time you call the package.
For example, I have a package called pk_mail. The package body looks like this:
CREATE OR REPLACE PACKAGE BODY     pk_mail
AS
mail_host_name     VARCHAR2 (30);
PROCEDURE   add_recipient    -- First procedure
END  send_to_all;            -- Last procedure
-- The following is run the first time the package is used in a given session:
BEGIN
     SELECT     txt
     INTO     mail_host_name
     FROM     constant_tbl
     WHERE     ent_name     = 'MAIL_HOST_NAME'
     AND     status_txt     = 'ACTIVE'
     AND     ROWNUM     = 1;
EXCEPTION
     WHEN     NO_DATA_FOUND
     THEN
          mail_host_name := 'itd-smtp-out.foo.bar.com;             -- Default value
END     pk_mail;
/The package variable mail_host_name, which is used in various procedures in the package, is set the first time the package is used. After the first call, the code that gives it a value is not run again.
Edited by: Frank Kulash on Jul 6, 2012 1:21 PM
Added pk_mail example.