know about and are free - oracledoug.comoracledoug.com/graham_wood_hidden_features.pdf ·...

Hidden Features You Probably Don’t

Know About And Are FREEGraham Wood

Oracle Corp

Features You Probably Don’t Know

About And Are MOSTLY FREE

Database Vault

Transparent Data Encryption

Grid Computing

Self Managing Database

XML Database

Oracle Data Guard

Real Application Clusters

Flashback Query

Virtual Private Database

Built in Java VM

Partitioning Support

Built in Messaging

Object Relational Support

Multimedia Support

Data Warehousing Optimizations

Parallel Operations

Distributed SQL & Transaction Support

Cluster and MPP Support

Multi-version Read Consistency

Client/Server Support

Platform Portability

Commercial SQL Implementation

1977

Oracle Database30 Years of Sustained Innovation

2007

4

Things that you might not know about

• SQL trace

5


DBMS_MONITOR.SESSION_TRACE_ENABLE(

session_id IN BINARY_INTEGER DEFAULT NULL,

serial_num IN BINARY_INTEGER DEFAULT NULL,

waits IN BOOLEAN DEFAULT TRUE,

binds IN BOOLEAN DEFAULT FALSE,

plan_stat IN VARCHAR2 DEFAULT NULL);

6


DBMS_MONITOR.SESSION_TRACE_ENABLE(

session_id IN BINARY_INTEGER DEFAULT NULL,

serial_num IN BINARY_INTEGER DEFAULT NULL,

waits IN BOOLEAN DEFAULT TRUE,

binds IN BOOLEAN DEFAULT FALSE,

plan_stat IN VARCHAR2 DEFAULT NULL);

• Frequency at which we dump row source statistics.

Value should be 'NEVER', 'FIRST_EXECUTION'

(equivalent to NULL) or 'ALL_EXECUTIONS'.

7


• SQL trace

– Enabling

– Capture row source data for more executions

– Trace file contents

– CLOSE statements captured

– Previously this was missing time

– Cursor mapping broken (11gR1)

– Statement text printed on every execute

– Long cursor numbers (11gR2)

– Not so „human readable‟

8


• SQL trace

– Enabling








– Tkprof enhancements

– Prints SQL id and plan hash for each statement

– Reports First, Avg, Max from multiple row source dumps

9


• SQL trace

– Enabling








– Tkprof enhancements

– Prints SQL id and plan hash for each statement

– Reports First, Avg, Max from multiple row source dumps

– But not CLOSE … doh!

10


• SQL trace

– Enabling trace for individual SQL statements

alter session set events 'sql_trace [sql: g3yc1js3g2689 |

7ujay4u33g337]'

11


• SQL

– Getting SQL info without making problem worse

V$SQLstat

12


• SQL

– Get outline data for a single statement

SELECT * FROM table (

DBMS_XPLAN.DISPLAY_CURSOR

('atfwcg8anrykp', NULL, „ADVANCED'));

– Get Estimated and Actual rows


DBMS_XPLAN.DISPLAY_CURSOR

(FORMAT=>'ALLSTATS LAST'));

Things that you might not know aboutSELECT /*+ gather_plan_statistics */ p.prod_name, SUM(s.quantity_sold)

FROM sales s, products p

WHERE s.prod_id =p.prod_id GROUP By p.prod_name ;


DBMS_XPLAN.DISPLAY_CURSOR(FORMAT=>'ALLSTATS LAST'));

14


• SQL

– Set a parameter for running an individual statement

OPT_PARAM hint

SELECT /*+

OPT_PARAM('star_transformation_enabled„,'true') */ *

FROM …..

15


• SQL

– Cause all statements to run without hints

_optimizer_ignore_hints=TRUE

– Extremely useful for testing

16


• Optimizer Statistics

– Pending Statistics

– Separates stats collection from stats publishing

– Capture stats as Pending

– Test using pending stats

– Publish once verified

Pending Statistics

• Controlled by parameter

• optimizer_use_pending_statistics

• Determines if the optimizer will use pending statistics

• Set to false by default in 11gR1

• Set at session level for testing

• Use dbms_stats package

• set_table_prefs

• All tables preferences have “publish” set to true by default

• publish_private_stats

• Once stats have been tested publish them for general use

• Monitor

• Look at dictionary table user_*_pending_stats (* = tab, col, ind)

18


• Optimizer Statistics

– Pending Statistics

– Separates stats collection from stats publishing

– Capture stats as Pending

– Test using pending stats

– Publish once verified

– Setting Statistics

– Use dbms_stats.set_*_stats

– Very useful for min/max values

– Be careful!

Setting Statistics

declare

minmax dbms_stats.numarray;

da dbms_stats.datearray;

sr dbms_stats.statrec;

begin

/* prepare min/max values */

/* date type column example */

da := dbms_stats.datearray

(to_date('01/01/1999','MM/DD/YYYY'),sysdate);

sr.epc := 2;

sr.bkvals := null;

dbms_stats.prepare_column_values(sr, da);

dbms_stats.set_column_stats('scott','emp','hiredate',srec=>sr);

end; /


• More Analytics

• Long awaited LISTAGG

• And many more….

AnalyticsSQL> select deptno,

2 listagg( ename, '; ' )

3 within group

4 (order by ename) enames

5 from emp

6 group by deptno

7 order by deptno

8 /

DEPTNO ENAMES

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

10 CLARK; KING; MILLER

20 ADAMS; FORD; JONES;

SCOTT; SMITH

30 ALLEN; BLAKE;

JAMES; MARTIN;

TURNER; WARD

Analytics

SQL> select deptno,

2 ename,

3 row_number()

4 over (partition by deptno

5 order by ename) rn,

6 first_value(ename)


8 order by ename) "1st ename",

9 nth_value(ename,3)


11 order by ename) "3rd ename",

12 last_value(ename)


14 order by ename

15 rows between current row

16 and unbounded following) "last ename"

17 from emp

18 order by deptno, ename

19 /

Analytics

SQL> select deptno,

2 ename,

3 row_number()


5 order by ename) rn,

6 first_value(ename)


8 order by ename) "1st ename",

9 nth_value(ename,3)


11 order by ename) "3rd ename",

12 last_value(ename)


14 order by ename

15 rows between current row

16 and unbounded following) "last ename"

17 from emp

18 order by deptno, ename

19 /

DEPTNO ENAME RN 1st e 3rd ena last en

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

10 CLARK 1 CLARK MILLER

KING 2 CLARK MILLER

MILLER 3 CLARK MILLER MILLER

20 ADAMS 1 ADAMS SMITH

FORD 2 ADAMS SMITH

JONES 3 ADAMS JONES SMITH

SCOTT 4 ADAMS JONES SMITH

SMITH 5 ADAMS JONES SMITH

30 ALLEN 1 ALLEN WARD

BLAKE 2 ALLEN WARD

JAMES 3 ALLEN JAMES WARD

MARTIN 4 ALLEN JAMES WARD

TURNER 5 ALLEN JAMES WARD

WARD 6 ALLEN JAMES WARD


• External tables allow for a preprocessor

– Program is run when you SELECT from external table

– The „location‟ is passed to the script/executable

– The executable does whatever it wants and writes to stdout

– Stdout is treated as the input file

– Most common use: gzip

External Table PreprocessorSQL> CREATE TABLE EMP_ET

2 (

3 "EMPNO" NUMBER(4),

4 "ENAME" VARCHAR2(10),

5 "JOB" VARCHAR2(9),

6 "MGR" NUMBER(4),

7 "HIREDATE" DATE,

8 "SAL" NUMBER(7,2),

9 "COMM" NUMBER(7,2),

10 "DEPTNO" NUMBER(2)

11 )

12 ORGANIZATION external

13 ( TYPE oracle_loader

14 DEFAULT DIRECTORY load_dir

15 ACCESS PARAMETERS

16 ( RECORDS DELIMITED BY NEWLINE

17 preprocessor exec_dir:'run_gunzip.sh'

18 FIELDS TERMINATED BY "|" LDRTRIM

19 )

20 location ( 'emp.dat.gz')

21 )

22 /

Table created.

External Table Preprocessor

SQL> !file emp.dat.gz

emp.dat.gz: gzip compressed data, was "emp.dat", from Unix, last

modified: Wed Oct 7 12:48:53 2009

SQL> !cat run_gunzip.sh

#!/bin/bash

/usr/bin/gunzip -c $*

SQL> select empno, ename from emp_et where rownum <= 5;

EMPNO ENAME

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

7369 SMITH

7499 ALLEN

7521 WARD

7566 JONES

7654 MARTIN

27


• Other stuff

– Out of line upgrade

– Full install rather than upgrade of $ORACLE_HOME

– 11.2.0.2

28


• Other stuff

– Connecting to a Hung Database

– Primarily to get diagnostic etc

Sqlplus –prelim / as sysdba

Not FREE, unless you have the correct

options

30


• SQL

– Change number of SQL captured by AWR

dbms_workload_repository.modify_snapshot_settings

(topnsql=> 200)

31


• SQL

– Always capture a statement even if not high load

dbms_workload_repository.add_colored_sql(:sql_id)

32


• Reporting

– AWR SQL Report

– Report history of execution of a SQL statement

– Plans in use during different time periods

– AWR Diff Report

– Highlights differences between two periods of AWR data

– AWR RAC Report

– Reports on distribution of workload across instances

– ASH Report

– Any time period

– Instance or filtered

– Filtering at sessions level almost equivalent of SQL trace

33


• Reporting

– AWR SQL Report

– Report history of execution of a SQL statement

– Plans in use during different time periods

– AWR Diff Report

– Highlights differences between two periods of AWR data

– AWR RAC Report

– Reports on distribution of workload across instances

– ASH Report

– Any time period

– Instance or filtered

– Filtering at sessions level almost equivalent of SQL trace

AND DON‟T FORGET THE ADDM REPORT!

34


• Reporting

– See ASH data in EM

– SQL Monitor

See Raw ASH data for Session in EM

Check Cardinality using SQL Monitor

SQL Monitor allows you to compare the estimated number of rows returned for each operation in the plan to actual rows returned


• Dumping ASH to file

Sqlplus –prelim / as sysdba

>oradebug setmypid

>oradebug dump ashdump 10

>alter session set events 'immediate

trace name ashdump level 10';

• 10 ==> minutes of history you want to dump

• Generated file can be loaded into database (or Excel)


• Dumping ASH to file (normal operation)

Sqlplus / as sysdba

>create table my_ash_data as select *

from v$active_session_history;

exp my_ash_data

• Load into another database

• Run custom queries

For More Information

http://search.oracle.com

or

www.oracle.com/database

oracle database 11g

know about and are free - oracledoug.comoracledoug.com/graham_wood_hidden_features.pdf ·...

Documents