Friday, February 17, 2017

Hotsos Symposium 2017 Feb27-Mar2

Here we go!  It’s just over a week away and it’s another excellent line up of speakers.   I’ll mention a few here to whet your appetite and go to our page on the line up to read more on these and the other great speakers. 

We have Larry Carpenter back for an excellent Keynote on Data Guard and the Cloud.   I’ve interested to hear what kind of cloud Data Guard works best in, Cirrus, Status, Cumulus….

Also coming all the way from down under again is Richard Foote, Mr. Oracle Index.  He’s talks last year were a huge hit as I’m sure they will be again this year.

Bryn Llewellyn will be back to spread the word about EBR. He’s also supporting his partner in crime Toon Koppelaars. Toon will take up where Bryn’s “Why Use PL/SQL” left off last year to present a real-world comparison of the NoPlsql & ThickDB paradigms. 

Toon and Vlado Braun will be doing the training day (Thursday March 2nd).  They will bring to the table their vast experience on Real-World Performance Training.

We have Jim Czuprynski to talk on In-Memory and some new Analytics in 12cr2, that’s going to be supper cool stuff.  These analytics and in-memory features are taking SQL statements to warp speed, really great stuff.

Carlos Sierra with be there taking in depth about SQL Trace and execution plans for beginners, but knowing Carlos, his talk will be valuable for all anyone.  Be you a wet-behind-the-ears newbie or an old seasoned veteran.

Paul Jackson brings shares with us his expertise on using SQL Monitor to do tuning, if you’ve not used SQL Monitor yet, you need to see this.  Even if you have Paul’s likely got so cool tricks and techniques you want to catch.

Kerry Osborne is back again with big talk on a big data.  Kerry’s tell it like it is style is sure to cut to the core of what this big data thing is all about and the how to use it with lots of technical discussion and not just market fluff.

Tim Gorman brings in an excellent discussion on testing with Data Virtualization, being from Colorado I’m wondering what this type of “virtualization” this will be.  I’ll have to make sure I get there to get the full story.

Many others will be talking as well; I hope this short list of a few of the folks speaking peaks your interest to attend.   If you haven’t registered there is still time!  There is a virtual option as well, but it’s always best to be there so you can mingle with the speakers and other oracle professionals to talk the talk so you can walk the walk.

Go here to Register:

Wednesday, February 1, 2017

Advanced Format option in DISPLAY_CURSOR

Oracle is good at sneaking things it to the database.  And the Advanced format option for DISPLAY CURSOR is one of those things.   Using by itself you get the Basic plan, the Alias information,  The Outline data,  Predicate information, column projection information and notes. 

The cool stuff is the outline data.  For those of us who have been around Oracle for a while this is the old “stored outline” which is the set of hints that would make a sorted outline.  Today these are the hints used to make a baseline.  They can just be cool to read and see what the hints would be to create the plan and to give you some insight into the hints themselves.

Here a quick example.  First the little script I used in SQLcl:  (Yea I’m moving to SQLcl, really a great tool.  A big step up for out command line tool, SQL*Plus is still great and a wonderful tool, SQLcl take all the greatness of SQL*Plus and adds some great functionality.)

set serveroutput off

create or replace view orderview as
select /*+ qb_name(orderview) */
  ord2.cust_no custnum, ord2.order_no ordnum,
  order_date orddate, product_id prodid, quantity
from ord_item2, ord2
where ord_item2.order_no = ord2.order_no

set termout off
select /*+ qb_name(mainblock) */lastname, custnum, ordnum, orddate, prodid, quantity
from customer2 c2, orderview ov
where ov.custnum = c2.cust_no;
set termout on

select * from table(dbms_xplan.display_cursor(format=>'ADVANCED'));

 Here’s the run:

SQL> @rvd02
SQL_ID  bsu4vh744stwq, child number 0
select /*+ qb_name(mainblock) */lastname, custnum, ordnum, orddate,
prodid, quantity from customer2 c2, orderview ov  where ov.custnum =

Plan hash value: 4000018549

| Id  | Operation           | Name      | Rows  | Bytes | Cost (%CPU)| Time     |
|   0 | SELECT STATEMENT    |           |       |       |   118 (100)|          |
|*  1 |  HASH JOIN          |           | 70975 |  3881K|   118   (1)| 00:00:01 |
|*  2 |   HASH JOIN         |           | 12890 |   553K|    51   (0)| 00:00:01 |
|   3 |    TABLE ACCESS FULL| CUSTOMER2 |  1000 | 26000 |     7   (0)| 00:00:01 |
|   4 |    TABLE ACCESS FULL| ORD2      | 12890 |   226K|    44   (0)| 00:00:01 |
|   5 |   TABLE ACCESS FULL | ORD_ITEM2 | 70975 |   831K|    66   (0)| 00:00:01 |

Query Block Name / Object Alias (identified by operation id):

   1 - SEL$19724973
   3 - SEL$19724973 / C2@MAINBLOCK
   4 - SEL$19724973 / ORD2@ORDERVIEW
   5 - SEL$19724973 / ORD_ITEM2@ORDERVIEW

Outline Data

      FULL(@"SEL$19724973" "C2"@"MAINBLOCK")
      FULL(@"SEL$19724973" "ORD2"@"ORDERVIEW")
      FULL(@"SEL$19724973" "ORD_ITEM2"@"ORDERVIEW")
      LEADING(@"SEL$19724973" "C2"@"MAINBLOCK" "ORD2"@"ORDERVIEW"
      USE_HASH(@"SEL$19724973" "ORD2"@"ORDERVIEW")
      USE_HASH(@"SEL$19724973" "ORD_ITEM2"@"ORDERVIEW")

Predicate Information (identified by operation id):

   1 - access("ORD_ITEM2"."ORDER_NO"="ORD2"."ORDER_NO")
   2 - access("ORD2"."CUST_NO"="C2"."CUST_NO")

Column Projection Information (identified by operation id):

   1 - (#keys=1) "ORD2"."ORDER_NO"[NUMBER,22],
       "ORD2"."CUST_NO"[NUMBER,22], "LASTNAME"[VARCHAR2,20],
   2 - (#keys=1) "ORD2"."CUST_NO"[NUMBER,22], "LASTNAME"[VARCHAR2,20],
   3 - (rowset=200) "C2"."CUST_NO"[NUMBER,22], "LASTNAME"[VARCHAR2,20]
   4 - (rowset=200) "ORD2"."ORDER_NO"[NUMBER,22],
       "ORD2"."CUST_NO"[NUMBER,22], "ORDER_DATE"[DATE,7]

   - this is an adaptive plan