DBMS_PROFILER

Similar Questions

• Steps using the DBMS_PROFILER package.

  Hi all

  I'd like to collect statistics for my stored procedure using DBMS_PROFILER package, but when I run the package by using the sample below, I get the following errors. I read the Oracle documentation, you need to have some loaded scripts (PROFTAB.sql.) etc. to execute it.

  Could someone, please give me the exact measurements that are necessary, if I can use this package in my diagram. Not as with which the user scripts should be executed, if scripts must be run in my schema, or etc., to grant all privileges and so on.

  Thank you.

  Below is executed the sample and the error I get. I use the version of Oracle 11 g.

  Samples:

  BEGIN

  DBMS_PROFILER.start_profiler ('test_profile');

  test;

  DBMS_PROFILER.stop_profiler;

  END;

  Error report-

  ORA-06528: PL/SQL Profiler generator runtime error

  ORA-06512: at "SYS." DBMS_PROFILER", line 123

  ORA-06512: at "SYS." DBMS_PROFILER", line 132

  ORA-06512: at line 2

  06528 00000 - "PL/SQL runtime error profile.

  * Cause: An error occurred during the execution of a PL/SQL Profiler

  operating mode.

  ORACLE-BASE - Oracle9i DBMS_PROFILER

  See you soon,.

  Manik.

• DBMS_PROFILER need education

  Hello

  DECLARE
    l_result  BINARY_INTEGER;
  BEGIN
  --  l_result := DBMS_PROFILER.FLUSH_DATA;
    l_result := DBMS_PROFILER.start_profiler
    (run_comment => 'do_something: ' || SYSDATE);
    wedb.testproc;
    l_result := DBMS_PROFILER.stop_profiler;
  END;
  /

  I have used the code above to learn the features of DBMS_PROFILER, learned something.
  Then I tried to see the code PACKAGE SYS. DBMS_PROFILER, I see the spec but not the body. Please advice if ORACLE usually encapsulates this to the top of all users.
  I also question below
  -In the above code when I used "l_result: = DBMS_PROFILER.start_profiler (run_comment = >" foo: "|"). SYSDATE); ", which the Sub part of the spec DBMS_PROFILER gets used?
  
  FUNCTION start_profiler)
  run_comment IN VARCHAR2: = SYSDATE,.
  run_comment1 IN VARCHAR2: = ".
  run_number ON the directory
  )
  DIRECTORY BACK;
  
  PROCEDURE start_profiler)
  run_comment IN VARCHAR2: = SYSDATE,.
  run_comment1 IN VARCHAR2: = ".
  run_number ON the directory
  );
  
  FUNCTION start_profiler)
  run_comment IN VARCHAR2: = SYSDATE,.
  run_comment1 IN VARCHAR2: = "
  )
  DIRECTORY BACK;
  
  PROCEDURE start_profiler)
  run_comment IN VARCHAR2: = SYSDATE,.
  run_comment1 IN VARCHAR2: = "
  );
  
  Thank you

  You use "l_result: = DBMS_PROFILER.start_profiler (run_comment => ' foo: ' |)» SYSDATE); ». It is an assignment (http://download.oracle.com/docs/cd/B28359_01/appdev.111/b28370/assignment_statement.htm#LNPLS01301)

  If you look at the link, you will see that after ": =" symbol, you can have as an expression. Instead of an expression, you can have a function. Instead of a statement, you can have a procedure. If your use exclude procedures, but the leaves open two functions.

  If a function has a parameter out (you better would not have an ever, but that's another debate), then you must specify one in the call.
  You do not have, so the only option left is the use of the third: the function without parameters.

  I hope this helps.

  Kind regards
  Rob.

• How to configure the parameters of coverage of codes for SQL / PL - SQL with TeamCity or Jenkins? It will be very useful for me in my project.

  Hello

  I need to set up a process of continuous integration for my oracle database. We plan to use Team City and SVN as source control tool.  For the unit test framework, we'll add utPLSQL. But we must see a report of the code coverage for the test cases. When I Googled it I for additional information, the cone for utPLSQL does not provide for the code coverage with Jenkins, while ruby-plsql has provide. I need help in the implementation of the code coverage for my test in PL SQL framework.  Based on your contributions, we will develop a model of POC and if it works perfectly, we'll use this practice to the coverage across the tests. Righ now we have two options to go to the establishment of CI process

  (1) City team (Servers0 CI = > Source Control (SVN) == > build (Ant scripts) == > deploy == > Testing Framework (utPLSQL) = > Code Coverage (?) = > publish the results of the team by email)

  (2) Jenkins = > Source Control (SVN) = > Build = > deployment = > Testing Framework (utPLSQL) == > = Code coverage > publish the result of the cover and test code to the team.

  UTPLSQL | Oracle's thoughts

  Therefore, choosing a Code coverage tool that is utPLSQL and integrate also with Jenkins or Teamcity. Those who have already worked on this part, please do not share your contributions and suggestions.

  Thank you

  Jean Yves

  Hello

  Basically, we take the exit of dbms_profiler and generate code coverage statistics.

• ORA-06528: PL/SQL Profiler generator runtime error

  Version 11.2.0.3.  I have installad DBMS_PROFILER:

  SYS AS SYSDBA > @?/rdbms/admin/profload.sql

  SYS AS SYSDBA > @?/rdbms/admin/proftab.sql

  
  

  I tried to use dbms_profiler.start_profiler in PLSQL code with user NUFN and discovered that lacked the NUFN user privileges on the objects of the Profiler. So I ran the following grants:

  grant the update, insert and delete on plsql_profiler_units to NUFN;

  grant the update, insert and delete on plsql_profiler_runs to NUFN;

  Grant execute on DBMS_PROFILER to NUFN;

  But still no luck:

  SQL > START

  dbms_profiler.start_profiler ('test1');

  EXECUTE IMMEDIATE 'select max (STATS_UPDATE_TIME), owner of the DBA_TAB_STATS_HISTORY group from owner to owner.

  dbms_profiler.stop_profiler ();

  END;

  /

  BEGIN

  *

  ERROR on line 1:

  ORA-06528: PL/SQL Profiler generator runtime error

  ORA-06512: at "SYS." DBMS_PROFILER", line 123

  ORA-06512: at "SYS." DBMS_PROFILER", line 132

  ORA-06512: at line 2

  I can't find what is missing from my user NUFN for her to be able to run DBMS_PROFILER.  Help, please!

  Thanks Bella, I had not run @?/rdbms/admin/proftab.sql as my user NUFN but as sys.  But you know what?  Even after execution of @?/rdbms/admin/proftab.sql as NUFN, the problem persisted!

  So I traced my session, read the trace file and found nothing wrong (but I have no experience in finding fault with the workings of dbms_profiler).  But I found the SQL code executed by the dbms_profiler.start_profiler command, which is:

  PARSING IN CURSOR #4574691968 len = 155 dep = 2 uid = 155 oct 2 lid = 155 = tim = hv 530047132716 = ad 3718124591 = "7000000946da848" sqlid = "akx2pxvftw51g."

  INSERT INTO PLSQL_PROFILER_RUNS (RUNID, RUN_DATE, RELATED_RUN, RUN_OWNER) VALUES (PLSQL_PROFILER_RUNNUMBER. NEXTVAL,: B1,: B2, USER) RETURN RUNID IN: O0

  END OF STMT

  ANALYSIS #4574691968:c = 315, e = 844, p = 0, cr = 0, cu = 0, put = 1, r = 0, dep = 2, og = 1, plh = 0, tim = 530047132715

  I tried manually the same HOME, and bingo, I discovered the cause of my problem:

  INSERT INTO PLSQL_PROFILER_RUNS (RUNID, RUN_DATE, RELATED_RUN, RUN_OWNER) VALUES (PLSQL_PROFILER_RUNNUMBER. NEXTVAL, sysdate, 'manually', 'ME');

  ORA-01950: no privileges on tablespace execution times

  After performing ALTER USER NUFN QUOTA UNLIMITED ON STATES; the problem has been resolved.

  Conclusion: after DBMS_PROFILER installation for a user, make sure that the user can write in objects DBMS_PROFILER (SUBSIDIES and QUOTAS).

• package pl/sql execution plan

  Hi all

  How we can find the bottoleneks in the pl/sql package. I know we can use dbms_profiler. Gathersed data, but not having planned the details.

  Here's the query I use

  Select s.text,

  p.total_occur,

  p.TOTAL_TIME/1000000000 total_time,

  p.min_time / 1000000000 min_time,.

  p.max_time / 1000000000 max_time

  of plsql_profiler_data p, user_source s, plsql_profiler_runs r

  where p.line # = Check

  and p.runid = r.runid

  and r.run_comment = 'XXXX_YYY_INIT.XXX_TO_DDD(1,,360). '

  and s.name LIKE 'XXX % ';

  I see that dbms_profiler has gathered data. But it does give that which line is how long take.

  Please guide findout that.

  Thank you

  Col owner format a10

  Col name format a20

  row col 9999 format

  Col source format a40 word_wrap

  occurrences of col 9999999 format

  Col tot_time_s format 999,999,990.999

  Col run_comment format a40

  format of pct col 999

  Select runid, run_date, plsql_profiler_runs run_comment;

  Select OWNER u.unit_owner, u.unit_name NAME, d.line # LINE, (select the text from all_source where type in ('BODY of PACKAGE', 'FUNCTION', 'PROCEDURE', 'TRIGGER') and name = u.unit_name and = d.line line # and u.unit_owner and type = u.unit_type = owner) SOURCE, d.total_occur OCCURRENCES, (d.total_time/1000000000) TOT_TIME_S, d.total_time/r.run_total_time*100 "PCT".

  of plsql_profiler_runs r, plsql_profiler_units u, plsql_profiler_data d

  where are.runid = & runid

  and r.runid = u.runid

  and r.runid = d.runid

  and d.unit_number = u.unit_number

  and (d.total_time/1000000000) > 1 - more than 1 second

  order of tot_time_s

  /

  Kind regards

  Harman

• How to make a Plan to explain on PL - SQL?

  Hey guys, it's Xev, I'm back with a whole new positive attitude . I am very happy now that I got my cluster (11.2) 2-node RAC 11 GR 2 installed and working properly

  I'm doing a plan to explain some pl - sql, I wrote, but I'm having a hard time trying to figure out how to make it work.

  The catplan.sql has been performed and I a plan_table ready and waiting for me.

  Here's the full code as well as the two tables table that I use, just as I'm supposed to display.

  This is the first table

  and Yes, I only want a single folder in there...

  CREATE TABLE FILLER

  ("DATA_PACKED" VARCHAR2 (75 BYTE)

  )

  Insert in FILLING (DATA_PACKED) values ('KICKKKKKKKKK');

  commit;

  Then at the top on the list is the table of the HAMMER, I will insert all the data once the pl/sql program runs...

  This table is supposed to be empty...

  CREATE TABLE HAMMER

  ('H1' VARCHAR2 (100 BYTE),

  "H2" VARCHAR2 (100 BYTE)

  )

  and now, the CODE... Now, it is a very small version of my load of GR 11, 2 test scripts that I wrote, but I need to see execution plan behind this logic.

  I can even make a plan to explain on this code, or do I have to use something else?
  

  ---------------------------------------------------------------------------------------------------------

  declare

  Table type is table of the varchar2 (30) index directory.

  Table Z_STUFF.

  Start

  Select * BULK COLLECT into Z_STUFF of filling;

  because me in 1... 100

  loop

  for zoo in 1... Z_STUFF. Count

  loop

  Insert into HAMMER (H1, H2)

  values (Z_STUFF (zoo), Z_STUFF (zoo));

  end loop;

  end loop;

  end;

  /

  ----------------------------------------------------------------------------------------------------

  All comments are welcome, I just need to see what makes this logic behind the scenes...

  Thank you

  Xev.

  A query plan refers to an SQL statement.  A PL/SQL block can have 0 or more SQL statements.  It is, therefore, unnecessary to speak to generate a query plan for a PL/SQL block.  Your PL/SQL block happens to contain two different SQL statements.  You might get two different query plans, one for each statement.  Given how trivial statements are, however, I'm not sure that query plans would give you a lot of information.

  What is the problem you're trying to solve?  If you try to set the PL/SQL, you probably want to look into various profilers of PL/SQL.  For example, the DBMS_PROFILER package or the package DBMS_HPROF.

  Justin

• Question performance tuning...

  Hi all...
  
  Database: Oracle 9i
  
  Sorry for the vague question.
  I have a process that runs in "PROD" for (24 h) too high, which I am trying to tune.
  But the problem is that the same process in "DEV" only "15 min". There is a reason for that.
  "PROD" and "DEV" data are not synchronized.
  
  Table of one of the main process questions following records
  
  PROD-> record 340 million.

   partioned by monthly, have indexes on most of the "important" columns that queries are using.

  IN DEV-> record 26 million
  

  i cannot copy the "PROD" data to "DEV"

  I tried the following things.
  
  (1) verified "explain plan" for all queries that are in the process.
  explain the plan looks good for most applications. I mean most applications use "index scans" rather than "full table scan.
  I know it's not the only thing that I should. There are several other things that could give the issue. But the plan of the explain command is ""not really bad.
  
  (2) accrued dbms_profiler to verify the statement that re take longer. I'm able to find "some" who need some settings.
  

  By the way, my queries in the process will use hadly 1 or 2 partitions. i mean 1 or 2 months data

  How can i tune something which is running is ok in "DEV", but not in "PROD"?
  Sorry i didn't give much details about any other stuff.

  Please guide me some steps I can do to find out what is the cause of the problem.
  
  Please let me know if I need to post some explain plans or any other outputs.
  
  Thanks in advance

  (1) verified "explain plan" for all queries that are in the process.

  Are the execution of PROD or DEV plans?
  You could try to simulate PROD on DEV important statistics.
  http://www.Oracle-base.com/articles/Misc/cost-based-optimizer-and-database-statistics.php#transfering_stats
  But it would be something that your DBA should help you.

• Explain the plan of a function?

  Hi people;
  
  I am running Oracle 11 g 11.2.0.3.
  Is it possible to get the plan of "explaining" execution of a function?
  
  Thx for your knowledge, looks.

  Hello

  dbms_profiler is really simple, it does not really require a tutorial.

  dbms_profiler.start_profiler;
  
  dbms_profiler.stop_profiler;
  

  That's all. The only thing that can take you a bit is the display of the results
  tables of % plsql_. Google will give you many links to examples, for example this one:

  http://www.toadworld.com/knowledge/KnowledgeXpertforOracle/tabid/648/TopicId/PROF7/default.aspx

  Best regards
  Nikolai

• Under the execution of pl/sql

  Hi all
  
  How to find the lowest performance of pl/sql programs, such as the way which we find bad execution SQL of AWR/tkprof etc.?
  
  
  concerning
  Sunil

  You can use DBMS_PROFILER http://download.oracle.com/docs/cd/E11882_01/appdev.112/e16760/d_profil.htm#ARPLS039.

• pl/sql profiler - data not listed is not in the Profiler tables?

  Hi all - just have a little problem with the pl/sql Profiler.
  

  Database Version: Oracle Database 10g Enterprise Edition Release 10.2.0.5.0 - 64bi
  Database Version: PL/SQL Release 10.2.0.5.0 - Production

  I created a test procedure:
  SQL > create or replace procedure testme
  2 as
  3
  number of v_customer_count 4;
  t_number v_level_table 5;
  Number 6 v_std_code_id;
  7
  Start 8
  9
  10. select count (*)
  11 in v_customer_count
  12 customer_master;
  13
  14 select level
  collection in bulk 15
  16 in v_level_table
  17 of the double
  18 connection of level < = 5000;
  19
  20
  21 end;
  22.
  
  Created procedure.
  
  
  and I run the procedure as follows:
  

  SQL> declare
    2     run_id number;
    3  begin
    4  
    5    run_id := dbms_profiler.start_profiler(to_char(sysdate,'dd-mm-yyyy hh24:mi:ss'));
    6    
    7    testme;
    8    
    9    dbms_profiler.flush_data;
   10    dbms_profiler.stop_profiler;
   11  
   12    dbms_output.put_line(run_id);
   13  
   14  end;
   15  /
  2
  
  PL/SQL procedure successfully completed.
  
  Elapsed: 00:00:00.08

  but:
  

  SQL> select * from sys.plsql_profiler_runs
    2  /
  
  no rows selected
  
  Elapsed: 00:00:00.00
  SQL> select * from sys.plsql_profiler_data;
  
  no rows selected

  What Miss me?

  Run proftab.sql in your schema, also compile in debug mode before the race.

• Ampersand substitution in create or replace procedure statement

  Hi guys,.
  
  I wonder why my replacement ampersand works in a create or replace statement of the stored procedure.
  
  CREATE OR REPLACE PROCEDURE UPDATE_DIM_SALES AS
  
  
  Cancel the DEFINITION DimSales;
  Cancel the DEFINITION FactTable;
  SET DimSales = 'TESTTAB;
  SET FactTable = myfact;
  
  
  BEGIN
  *...*
  
  Error (5,20): PLS-00103: encountered the symbol "=" when expected as follows: =. (@ %; not null default range character)
  
  
  If I assign the value with: = I get the error "invalid table" thereafter for the statemnt of INSERTION:
  
  CREATE OR REPLACE PROCEDURE UPDATE_DIM_SALES AS
  
  
  Cancel the DEFINITION DimSales;
  Cancel the DEFINITION FactTable;
  SET DimSales: = ' x 2';
  SET FactTable: = ' x 1';
  
  
  BEGIN
  INSERT INTO & DimSales (column1, column2,...)
  
  Why not the ampersand substitution work in a stored procedure?

  No problem with the carpet. Display. But you don't need to create and drop them again. Simply create a time and refresh just before departure if necessary. I expect the creation and updating of points of view is fast comparibly. Also access to the seller_id should be indexed.

  Your original pl/sql block is slow due to several reasons. One might be that you agree to each line. Engage is an extremely slow process. He avoid as much as possible. Another part might be that you make all the changes of context of pl/sql. The obvious way to avoid this would be to BULK operations (bulk-select + Forall). It is not so difficult. The size of the collection must be limited to a lower amount.

  Best way would be to measure the performance of pl/sql parts using a tool such as DBMS_PROFILER. Then you can decide what part is slow and must be granted.

  A simple insert could be faster. Did you test that?

  example not tested

  INSERT /*+append */
  INTO &DimSales (K_Sales,REG,BVL,DS, VS,RS,GS,VK)
  (SELECT trim(leading '0' from RS||GS) ,REG, BVL,DS, VS,RS,GS,VK
  from &FactTable f
  join &testsales s on f.SellerNo =s.Seller_No;
  
  commit;
  

  You can run the other issues, but this is generally the fastest approach. Try to avoid the separate if possible. This requires a sort and will slow down the insert of a lot.

  Published by: Sven w. on November 30, 2010 17:10

• test the performance of my proc

  I'm trying to find how long it takes to insert.
  Y at - it another way to find the exact time needed to insert in milliseconds.
  
  Here is my sample code. and the exit time. Are there
  
  create or replace procedure perftest
  IS
  I int: = 0;
  int j: = 0;
  int ID;
  T1 timestamp;
  T2 timestamp;
  T3 varchar2 (30);
  Start
  I'm in 1.20
  LOOP
  Select systimestamp in double t1;
  Select seq.nextval in double id;
  INSERT INTO dept (Id, name) values (id, 'keep')
  j in 1.100
  LOOP
  INSERT INTO values (Id, name, deptid) emp (j, 'sample', id);
  END LOOP;
  Select systimestamp in double t2;
  Select (t2 - t1) in t3 double;
  dbms_output.put_line (' the value is:-' | t3);
  END LOOP;
  end;
  /
  
  
  the value is:- + 00:00:00.047000000 000000000
  the value is:- + 00:00:00.000000000 000000000
  the value is:- + 00:00:00.016000000 000000000
  the value is:- + 00:00:00.000000000 000000000
  the value is:- + 00:00:00.015000000 000000000
  the value is:- + 00:00:00.016000000 000000000
  the value is:- + 00:00:00.000000000 000000000
  the value is:- + 00:00:00.015000000 000000000
  the value is:- + 00:00:00.000000000 000000000
  the value is:- + 00:00:00.016000000 000000000
  the value is:- + 00:00:00.000000000 000000000
  the value is:- + 00:00:00.000000000 000000000
  the value is:- + 00:00:00.016000000 000000000
  the value is:- + 00:00:00.000000000 000000000
  the value is:- + 00:00:00.015000000 000000000
  the value is:- + 00:00:00.000000000 000000000
  the value is:- + 00:00:00.016000000 000000000
  the value is:- + 00:00:00.016000000 000000000
  the value is:- + 00:00:00.000000000 000000000
  the value is:- + 00:00:00.000000000 000000000

  Install the package dbms_profiler.

  Add the Profiler command to your procedure.

  create or replace procedure perftest
  IS
  i int := 0;
  j int := 0;
  id int;
  t1 timestamp;
  t2 timestamp;
  t3 varchar2(30);
  RESULT NUMBER ;
  begin
  RESULT:=DBMS_PROFILER.START_PROFILER('prog');
  for i in 1..20
  LOOP
  select systimestamp into t1 from dual;
  select seq.nextval into id from dual;
  INSERT INTO dept (Id,Name) values (id,'store')
  for j in 1..100
  LOOP
  INSERT INTO emp (Id,name,deptid) values (j,'sample',id);
  END LOOP;
  select systimestamp into t2 from dual;
  select (t2-t1) into t3 from dual;
  dbms_output.put_line ('the value is :- '|| t3);
  END LOOP;
  RESULT:=DBMS_PROFILER.STOP_PROFILER ;
  end;
  /
  

  then run your procedure
  Review the statistics in the tables of Profiler.

  SELECT   u.unit_owner || '.' || u.unit_name unit, line#,
           ROUND (d.total_time / 1000000000) total_time_s, d.total_occur tc,
           SUBSTR (s.text, 1, 40) stext
      FROM plsql_profiler_runs r,
           plsql_profiler_units u,
           plsql_profiler_data d,
           all_source s
     WHERE r.runid = u.runid
       AND d.runid = u.runid
       AND u.unit_number = d.unit_number
       AND s.owner = u.unit_owner
       AND s.TYPE = u.unit_type
       AND s.NAME = u.unit_name
       AND s.line = d.line#
  ORDER BY line#
  /
  

  Sanjay

• Optimized SQL in sql * more... .acts differently in pl/sql... Why?

  Hi all
  
  I was reading an ad, and there I found the below (beware I'm a dba level jr). My question is how can we optimize the sql inside pl/sql if his will act differently... as below mention... the SQL query can act differently (it may or may not) incorporated in a pl/sql block, let so say I have a select which I optimized in sql and now put this same sql in pl/sql and it is different... How can you optimize it... also I know dbms_profiler... any good links for her would be useful as well...
  
  The optimizer mode parameter is intended only for statements that are derived directly and not from PL/SQL statements. Therefore, if you have tested your SQL querymore or any other tool, then then joined the same in a PL/SQL (anonymous or stored routines) block, the query may or may not work the same! DML in PL/SQL statements are executed to CHOOSE mode (if statistical ALL_ROWS is present) defined at the session level default and OPTIMIZER_MODE does not influence besoin.*
  
  http://www.databasejournal.com/features/Oracle/article.php/3084581/Oracle-optimizer-moving-to-and-working-with-CBO-part-3.htm
  (see section - 8.4) SQL and PL/SQL

  user630084 wrote:

  The optimizer mode parameter is intended only for statements that are derived directly and not from PL/SQL statements. Therefore, if you have tested your SQL querymore or any other tool, then then joined the same in a PL/SQL (anonymous or stored routines) block, the query may or may not work the same! DML in PL/SQL statements are executed to CHOOSE mode (if statistical ALL_ROWS is present) defined at the session level default and OPTIMIZER_MODE does not influence besoin.*

  http://www.databasejournal.com/features/Oracle/article.php/3084581/Oracle-optimizer-moving-to-and-working-with-CBO-part-3.htm
  (see section - 8.4) SQL and PL/SQL

  I think that this does not apply to the latest versions of Oracle longer. The hidden parameter "_optimizer_mode_force" is true because 8.1.6 and, therefore, the recursive SQL in PL/SQL should follow the same OPTIMIZER_MODE as defined in the session.

  You can find a more detailed explanation on the website of Jonathan Lewis on this topic: http://www.jlcomp.demon.co.uk/faq/sqlplsql.html

  Kind regards
  Randolf

  Oracle related blog stuff:
  http://Oracle-Randolf.blogspot.com/

  SQLTools ++ for Oracle (Open source Oracle GUI for Windows):
  http://www.sqltools-plusplus.org:7676 /.
  http://sourceforge.NET/projects/SQLT-pp/

• Procedure measure execution time...

  Hello
  
  Is it possible to measure how long each a database procedure take? Is it possible for V$ views Oracle or performance as AWR report? On the other hand, is it possible to measure the duration of execution of SQL statements, but without using "set timing on."
  
  Best,
  tutus

  To measure how long a process takes to run, use DBMS_PROFILER.

  To measure how long an individual SQL takes to run, use SQLTRACE and analyze results TKPROF.