Twp Upgrading 10g To 11g What To Expect From Optimizer

There are several new initialization parameters that govern the optimizer and its new features in
Oracle Database 11g. Below are the details on the new parameters.

In Oracle Database 11g a new feature called Invisible indexes was introduced. This feature allows
a DBA to create an index on a table without affecting any execution plan and thus not impacting
the performance of the application. An invisible index cannot be used by the optimizer as an
access path to that table. In order to test such indexes as a potential access path, the DBA needs
to set OPTIMIZER_USE_INVISIBLE_INDEXES to TRUE (default FALSE) within a session, and
then execute any SQL statement that could benefit from using the index to test its performance.
If the index proves to be useful, the DBA can mark it visible; if it does not, it can be dropped
without impacting the performance of the application.

Traditionally, when optimizer statistics are gathered they are published (written) immediately to
the appropriate dictionary tables and begin to be used by the optimizer. However, in Oracle
Database 11g it is possible to gather optimizer statistics but not have them published
immediately. Instead of going into the usual dictionary tables the statistics are stored in pending
tables so that they can be tested before they are published. A DBA can test these pending
statistics by doing an alter session command to set
OPTIMIZER_USE_PENDING_STATISTICS to TRUE. Then they should run any SQL statements
they believe might be affected by the new statistics. By testing the statistics before they are
published, the DBA gets an opportunity to prevent plan regressions that result from incomplete
or incorrect statistics.
More information on pending statistics can be found in the new features section below.

In Oracle Database 11g a new feature called SQL Plan Management (SPM) has been introduced
to guarantees any plan changes that do occur lead to better performance. When
OPTIMIZER_CAPTURE_SQL_PLAN_BASELINES is set to TRUE (default FALSE) Oracle will
automatically capture a SQL plan baseline for every repeatable SQL statement on the system.
The execution plan found at parse time will be added to the SQL plan baseline as an accepted
plan.

2

More information on SPM can be found in the new features section below.

Each SQL statement captured in SPM has a SQL plan baseline, which contains one or more
known or verified execution plans. When OPTIMIZER_USE_SQL_PLAN_BASELINES is set to
TRUE (default) the optimizer will only use one of these known plans even if a different plan is
found during SQL-compilation. This guarantees that any plan change is verified to have better
performance before using it.

The
ANALYZE In Oracle 8i a new PL/SQL package, called DBMS_STATS was introduced to gather and manage
command optimizer statistics. DBMS_STATS is Oracle‟s preferred method for gathering statistics. The
has been DBMS_STATS package has been extended in Oracle Database 11g to accommodate new types of
officially
statistics and monitoring data that can now be collected. Changes have also been made to the
obsolete
for
automatic statistics-gathering job that is enabled by default in 11g.
gathering
statistics.
Use the
DBMS_ST
ATS
package In previous releases you had to use the DBMS_STATS.SET_PARM procedure to change the default
instead. value for the parameters used by the DBMS_STATS.GATHER_*_STATS procedures. The scope
of any changes that were made was all subsequent operations. In Oracle Database 11g, the

3

DBMS_STATS.SET_PARAM procedure has been deprecated and it has been replaced with a set of
procedures that allow you to set a preference for each parameter at a table, schema, database, and
global level. These new procedures are called DBMS_STATS.SET_*_PREFS and offer a much
finer granularity of control. The list of parameters you can change are as follows:

AUTOSTATS_TARGET (SET_GLOBAL_PREFS only)
CASCADE
DEGREE
ESTIMATE_PERCENT
METHOD_OPT
NO_INVALIDATE
GRANULARITY
PUBLISH
INCREMENTAL
STALE_PERCENT

The SET_TABLE_PREFS procedure allows you to change the default values of the parameters
used by the DBMS_STATS.GATHER_*_STATS procedures for the specified table only.

The SET_SCHEMA_PREFS procedure allows you to change the default values of the parameters
used by the DBMS_STATS.GATHER_*_STATS procedures for all of the existing tables in the
specified schema. This procedure actually calls SET_TABLE_PREFS for each of the tables in the
specified schema. Since it uses SET_TABLE_PREFS calling this procedure will not affect any new
objects created after it has been run. New objects will pick up the GLOBAL preference values
for all parameters.

The SET_DATABASE_PREFS procedure allows you to change the default values of the parameters
used by the DBMS_STATS.GATHER_*_STATS procedures for all of the user-defined schemas in
the database. This procedure actually calls SET_TABLE_PREFS for each of the tables in each of
the user-defined schemas. Since it uses SET_TABLE_PREFS this procedure will not affect any
new objects created after it has been run. New objects will pick up the GLOBAL preference
values for all parameters. It is also possible to include the Oracle owned schemas (sys, system,
etc) by setting the ADD_SYS parameter to TRUE.

The SET_GLOBAL_PREFS procedure allows you to change the default values of the parameters
used by the DBMS_STATS.GATHER_*_STATS procedures for any object in the database that does
not have an existing table preference. All parameters default to the global setting unless there is a
table preference set or the parameter is explicitly set in the GATHER_*_STATS command.
Changes made by this procedure will affect any new objects created after it has been run. New
objects will pick up the GLOBAL_PREF values for all parameters. With GLOBAL_PREFS it is
also possible to set a default value for one additional parameter, called AUTOSTAT_TARGET. This

4

additional parameter controls what objects the automatic statistic gathering job (that runs in the
nightly maintenance window) will look after. The possible values for this parameter are ALL,
ORACLE, and AUTO. The default value is AUTO.

DBMS_STATS.GATHER*_STATS obeys the following hierarchy for parameter values; parameter
values explicitly set in the command overrule everything. If the parameter has not been set in the
command, we check for a table level preference. If there is no table preference set, we use the
GLOBAL preference. For example, if you wanted to switch off histogram creation for the SALES
table in the SH schema you can use:

BEGIN
DBMS_STATS.SET_TABLE_PREFS(‘SH’,'SALES','METHOD_OPT',’FOR ALL COLUMN SIZE
1);
END;

Very often with partitioned tables, a new empty partition will be added to the table and data will
begin to be loaded into it immediately. If the newly loaded data is queried before statistics can be
gathered for this partition, then the Optimizer will have to prorate its cardinality estimates for
these queries. Prorated cardinality estimates can lead to sub-optimal plans. By using
DBMS_STATS.COPY_TABLE_STATS procedure, it is possible to copy the statistics from one of
the other partitions in the table to the new partition. Column statistics (min, max, NDV,
histogram, etc), partition statistics (number of rows, blocks, etc) and statistics for local indexes
will be copied. The minimum and maximum values for the partitioning columns will be adjusted
to reflect the correct values for the new partition. In the following example, SALES_Q3_2000 is
the source partition and SALES_Q4_2000 is the target partition.
BEGIN
DBMS_STATS. COPY_TABLE_STATS (‘SH’,’SALES’,'SALES_Q3_2000',
'SALES_Q4_2000', FORCE=>TRUE);
END;

In real-world data, there is often a relationship or correlation between the data stored in different
columns of the same table. For example, in the CUSTOMERS table, the values in the
CUST_STATE_PROVICE column are influenced by the values in the COUNTRY_ID column, as the
state of California is only going to be found in the United States. Until now, the Optimizer had
no real way of knowing about these real-world relationships and could potentially miscalculate
the selectivity if multiple columns from the same table are used in the where clause of a
statement. With extended statistics you now have an opportunity to tell the Optimizer about
these real-world relationships.
By creating statistics on a group of columns, the Optimizer can be given a more accurate
selectivity guideline for the columns used together in a where clause of a SQL statement. Not all

5

of the columns in the column group need to be present in the SQL statement for the Optimizer
to use extended statistics; only a subset of the columns is necessary. Use
DBMS_STATS.CREATE_EXTENDED_STATS to define the column group you want to have statistics
gathered on as a whole. Once the group has been established Oracle will automatically maintain
the statistics on that column group when statistics are gathered on the table.
SELECT DBMS_STATS.CREATE_EXTENDED_STATS(null,'customers',
'(country_id, cust_state_province)');
FROM dual;

After creating the column group, you will see an additional column, with a system-generated
name, in user_tab_col_statistics table. This new column represents the column group.
It is also possible to create extended statistics for an expression (including functions), as it is
difficult for the Optimizer to estimate the cardinality of a where clause predicate that has
columns embedded inside expressions. For example, if it is common to have a where clauses
predicate UPPER(LastName)=:B1, then it would be beneficial to create extended statistics for
UPPER(LastName).

BEGIN
DBMS_STATS.GATHER_TABLE_STATS(null,'customers',method_opt =>
'for all columns size skewonly for columns(upper(cust_last_name))');
END;

Gathering statistics on partitioned tables consists of gathering statistics at both the table level and
partition level. Prior to Oracle Database 11g, adding a new partition or modifying data in a few
partitions required scanning the entire table to refresh table-level statistics. Scanning the entire
table can be very expensive as partitioned tables are generally very large. However, in Oracle
Database 11g this issue has been addressed with the introduction of incremental global statistics.
Typically with partitioned tables, new partitions are added and data is loaded into these new
partitions. After the partition is fully loaded, partition level statistics need to be gathered and the
global statistics need to be updated to reflect the new data. If the INCREMENTAL preference for
the partitioned table is set to TRUE, and the DBMS_STATS GRANULARITY parameter is set to
AUTO, Oracle will gather statistics on the new partition and update the global table statistics by
scanning only those partitions that have been added or modified and not the entire table. Below
are the steps necessary to use incremental global statistics.
BEGIN
DBMS_STATS.SET_TABLE_PREFS(‘SH’,’SALES’,‘INCREMENTAL’,’TRUE’);
END;

BEGIN
DBMS_STATS.GATHER_TABLE_STATS(‘SH’,’SALES’);
END;

6

Incremental Global Stats works by storing a synopsis for each partition in the table. A synopsis is
statistical metadata for that partition and the columns in the partition. Each synopsis is stored in
the SYSAUX tablespace and takes approximately 200KB. Global statistics are generated by
aggregating the synopses from each partition, thus eliminating the need to scan the entire table to
gather table level statistics (see Figure 1). When a new partition is added to the table you only
need to gather statistics for the new partition. The global statistics will be automatically updated
by aggregating the new partition synopsis with the existing partitions synopses.

When it comes to deploying a new application or application module it is standard practice to
test and tune the application in a test environment before it is moved to production. However,
even with testing it is possible that SQL statements in the application will have different
execution plans in production then they did on the test system. One of the key reasons an
execution plan can differ from one system to another (from test and production) is because the
optimizer statistics on each system are different. In Oracle Database 10g Release 2, the
DIFF_TABLE_STATS_* functions can be used to compare statistics for a table from two
different sources. The statistics can be from:
 A user statistics table and current statistics in the dictionary
 A single user statistics table containing two sets of statistics that can be identified using statids
 Two different user statistics tables
 Two points in history

7

 Current statistics and a point in history
 Pending Statistics with the current statistics in the dictionary
 Pending Statistics with a user statistics table
The function also compares the statistics of the dependent objects (indexes, columns, partitions).
The function displays statistics for the object(s) from both sources if the difference between the
statistics exceeds a specified threshold. The threshold can be specified as an argument to the
function; the default value is 10%. The statistics corresponding to the first source will be used as
the basis for computing the differential percentage.
In the example below, we compare the current dictionary statistics for the EMP table with the
statistics for EMP in the statistics table TAB1; the SQL statement will generate a report-like output
on the screen.
SQL> select report, maxdiffpct from
table(DBMS_STATS.DIFF_TABLE_STATS_IN_STATTAB(‘SCOTT’,’EMP’,’TAB1’ ));

In previous releases you could lock statistics for a table or a schema. Once statistics are locked,
no modifications can be made to those statistics until the statistics have been unlocked. In Oracle
Database 11g the DBMS_STATS package has two new procedures for locking and unlocking
statistics at the partition level. These additional procedures allow for a finer granularity of
control.
BEGIN
DBMS_STATS.LOCK_PARTITION_STATS(‘SH’,’SALES’, 'SALES_Q3_2000');
END;

As discussed in the previous section, it is now possible to gather optimizer statistics but not have
them published immediately. To activate pending statistics you need to use
DBMS_STATS.SET_*_PREFS procedure to change value of the parameter PUBLISH from TRUE
(default) to FALSE for the object(s) you wish to create pending statistics for.
BEGIN
DBMS_STATS.SET_TABLE_PREFS(‘SH’,’SALES’,‘PUBLISH’,’FALSE’);
END;

Then simple gather statistics on the object(s) as normal.
BEGIN
DBMS_STATS.GATHER_TABLE_STATS(‘SH’,’SALES’);
END;

The statistics gathered for these objects can be displayed using the dictionary views called
USER_*_PENDING_STATS. These statistics can be tested by issuing an alter session command to

8

set the initialization parameter OPTIMIZER_USE_PENDING_STATS to TRUE and running the
necessary queries. Once you are happy with the pending statistics you can publish them using the
new procedure PUBLISH_PENDING_STATS.
BEGIN
DBMS_STATS.PUBLISH_PENDING_STATS(‘SH’,’SALES’);
END;

It is critical to provide the Optimizer with accurate statistics and traditionally the most accurate
statistics are gathered when all rows in the table are processed. However, gathering statistics by
doing a full scan of larger tables is extremely time consuming and is rarely an option for most
customers. In fact most customers use sampling to gather statistics as it executes quickly. But
without detailed testing, it is difficult to know the sample size to use to get accurate statistics. In
the past, Oracle has tried to address this issue by introducing AUTO_SAMPLE_SIZE. With
AUTO_SAMPLE_SIZE Oracle would automatically determine the sample size to give good
statistics. However, AUTO_SAMPLE_SIZE had an Achilles heel; the sample size determined by
AUTO_SAME_SIZE may be inaccurate for computing statistics when there is an extreme skew in
the data. In Oracle Database 11g the sampling algorithm has been completely rewritten. The new
algorithm is hash based and provides deterministic statistics, which have the accuracy of
computing the statistics but with the speed of a 10% sample. This new algorithm is used when
estimate_percent is AUTO_SAMPLE_SIZE (default) in any of the
DBMS_STATS.GATHER_*_STATS procedures.

Oracle will automatically collect statistics for all database objects, which are missing statistics or
have stale statistics by running an Oracle AutoTask during a predefined maintenance window
(10pm to 2am weekdays and 6am to 2am at the weekends).
This AutoTask gathers optimizer statistics by calling the internal procedure
DBMS_STATS.GATHER_DATABASE_STATS_JOB_PROC. This procedure operates in a very similar
fashion to the DBMS_STATS.GATHER_DATABASE_STATS procedure using the GATHER AUTO
option. The primary difference is that Oracle internally prioritizes the database objects that
require statistics, so that those objects, which most need updated statistics, are processed first.
You can verify that the automatic statistics gathering job exists by viewing the
DBA_AUTOTASK_CLIENT_JOB view:
SELECT client_name, status
FROM DBA_AUTOTASK_TASK
WHERE client_name like 'auto optimizer %';

Statistics on a table are considered stale when more than STALE_PERCENT (default 10%) of the
rows are changed (total # of inserts, deletes, updates) in the table. Oracle monitors the DML

9

activity for all objects and records it in the SGA. The monitoring information is periodically
flushed to disk and is exposed in the *_tab_modifications view.

SELECT TABLE_NAME, INSERTS, UPDATES, DELETES
FROM USER_TAB_MODIFICATIONS;

It is also possible to manually flush this data by calling the procedure
DBMS_STATS.FLUSH_MONITORING_INFO.

The automatic statistics-gathering job uses the default parameter values for the DBMS_STATS
procedures. If you wish to change these default values you can use the
DBMS_STATS.SET_GLOBAL_PREFS procedure. Remember these values will be used for all
schemas including „SYS‟. To change the 'STALE_PERCENT' you can use

BEGIN
DBMS_STATS.SET_GLOBAL_PREFS('STALE_PERCENT','5');
END;

If you already have a well-established statistics gathering procedure or if for some other reason
you need to disable automatic statistics gathering altogether, the most direct approach is to
disable the GATHER_STATS_JOB as follows:

BEGIN
DBMS_AUTO_TASK_ADMIN.DISABLE(
client_name => 'auto optimizer stats collection',
operation => NULL,
window_name => NULL);
END;

If you choose to switch off the automatic statistics gathering job for your main application
schema, consider leaving it on for the dictionary tables. You can do this by changing the value of
AUTOSTATS_TARGET to ORACLE instead of AUTO using DBMS_STATS.SET_GLOBAL_PREFS.

Execution plan stability has always been somewhat of a holy grail in the optimizer space and
several features have been introduced in previous releases to improve it, such as Stored Outlines
and SQL Profiles. However, these methods would also prevent the optimizer from finding better
plans when data volume changes. In Oracle Database 11g plan stability has been addressed once
and for all with the introduction of SQL Plan Management (SPM).
SQL plan management (SPM) ensures that runtime performance will never degrade due to the
change of an execution plan. To guarantee this, only accepted (trusted) execution plans will be
used; any plan will be tracked and evaluated at a later point in time and only accepted as verified

10

if the new plan performs better than an accepted plan. SQL Plan Management has three main
components:
1. SQL plan baseline capture:
Create SQL plan baselines that represents accepted execution plans for all relevant SQL
statements. The SQL plan baselines are stored in a plan history inside the SQL
Management Base in the SYSAUX tablespace.
2. SQL plan baseline selection
Ensure that only accepted execution plans are used for statements with a SQL plan
baseline and track all new execution plans in the history for a statement as unaccepted
plan. The plan history consists of accepted and unaccepted plans. An unaccepted plan
can be unverified (newly found but not verified) or rejected (verified but not found to
performant).
3. SQL plan baseline evolution
Evaluate all unverified execution plans for a given statement in the plan history to
become either accepted or rejected.

11

For SPM to work, you must first seed the SQL Management Base with the execution plans,
which will become the SQL plan baseline for each statement. There are two different ways to
populate a SQL Management Base:
 Automatic capture
 Bulk load

Automatic capture can be switched on by setting the init.ora parameter
Capture
OPTIMIZER_CAPTURE_SQL_PLAN_BASELINES to TRUE (default FALSE). When automatic
plans “on
capture is enabled, the SPM repository will be automatically populated for any repeatable SQL
the fly “or
bulk load statement. To identify repeatable SQL statements, the optimizer will log the identity (SQL
SPM with Signature) of each SQL statement into a statement log the first time it is compiled. If the SQL
plans from statement is processed again (executed or compiled) the presence of its identity in the statement
the cursor log will signify it to be a repeatable statement. A SQL plan history will be created for the
cache, a
statement, which will include information used by the optimizer to reproduce the current
SQL
execution plan, such as the SQL text, outline, bind variables, and compilation environment. The
Tuning Set
or import current execution plan will be added as the first SQL plan baseline and this plan will be marked
plans from as accepted. Only accepted plans will be used; if some time in the future a new plan is found for
another this SQL statement, the execution plan will be added to the plan history and will be marked for
system. verification. It will only be marked accepted if its performance is better than that of a plan
chosen from the current SQL plan baseline.

Bulk loading of execution plans is especially useful when a database is being upgraded from a
previous version to Oracle Database 11g or when a new application is being deployed. Bulk
loading can be done in conjunction with, or instead of, automatic plan capture. Execution plans
that are bulk loaded are automatically used to create new SQL plan baselines or to add to an
existing one. The new SQL plan baselines are marked as accepted. The SQL Management Base
can be bulk loaded using four different techniques:
1. Populate the execution plans for a given SQL Tuning Set (STS)
2. Populate the execution plans from Stored Outlines
3. Use the execution plans currently in the Cursor Cache
4. Unpack existing SQL plan baselines from a staging table
More information on using bulk loading during upgrade can be found in the Preparing to
Upgrade Section.

12

Each time a SQL statement is compiled, the Optimizer first uses the traditional cost-based search
method to build a best-cost plan. If the initialization parameter
OPTIMIZER_USE_PLAN_BASELINES is set to TRUE (default) then before the cost based plan is
executed the optimizer will try to find a matching plan in the SQL statement‟s SQL plan baseline;
this is done as in-memory operation, thus introducing no measurable overhead to any
application. If a match is found then it proceeds with this plan. Otherwise, if no match is found,
the newly generated plan will be added to the plan history; it will have to be verified before it can
be accepted as a SQL plan baseline. Instead of executing the newly generated plan the optimizer
will cost each of the accepted plans for the SQL statement and pick the one with the lowest cost
(note that a SQL plan baseline can have more than one verified/accepted plan for a given
statement). However, if a change in the system (such as a dropped index) causes all of the
accepted plans to become non-reproducible, the optimizer will use the newly generated cost-
based plan.

Plans can
be manually When the optimizer finds a new plan for a SQL statement, the plan is added to the plan history
evolved or as a non-accepted plan that needs to be verified before it can become an accepted plan. It is
verified at
possible to evolve a SQL statement‟s execution plan using Oracle Enterprise Manager or by
any time or
running the command-line function DBMS_SPM.EVOLVE_SQL_PLAN_BASELINE. Using either of
you can
schedule a
these methods you have three choices:
database
1. Accept the plan only if it performs better than the existing SQL plan baseline
job to run
the evolve 2. Accept the plan without doing performance verification
process.
3. Run the performance comparison and generate a report without evolving the new plan.
If you choose option 1, it will trigger the new plan to be evaluated to see if it performs better
than a selected plan baseline. If it does, then the new plan will be added to the SQL plan baseline,
as an accepted plan. If not the new plan will remain in the plan history as a non-accepted plan
but its LAST_VERIFIED attribute will be updated with the current timestamp. A formatted text
report is returned by the function, which contains the actions performed by the function as well
as side-by-side display of performance statistics of the new plan and the original plan.
More information on SQL Plan Management can be found in the Preparing to Upgrade section.

13

Oracle introduced the bind peeking feature in Oracle 9i. With bind peeking, the Optimizer peeks
With at the values of user-defined bind variables on the first invocation of a cursor. This allows the
Adaptive
optimizer to determine the selectivity of any WHERE clause condition as if literals have been
Cursor
used instead of bind variables, thus improving the quality of the execution plan generated for
Sharing a
Statement
statements using bind variables.
with binds
can have However, there was a problem with this approach, when the column used in the WHERE clause
multiple with the bind contained a data skew. If there is data skew in the column, it is likely that a
execution
histogram has been created on this column during statistics gathering. When the optimizer peeks
plans
at the value of the user-defined bind variable and chooses a plan, it is not guaranteed that this
plan will be good for all possible values for the bind variable. In other words, the plan is
optimized for the peeked value of the bind variable, but not for all possible values.
In Oracle Database 11g, the optimizer has been enhanced to allow multiple execution plans to be
used for a single statement that uses bind variables. This ensures that the best execution plan will
be used depending on the bind value.
A cursor will be marked bind sensitive if the optimizer believes the optimal plan may depend on
the value of the bind variable. When a cursor is marked bind sensitive, Oracle monitors the
behavior of the cursor using different bind values, to determine if a different plan for different
bind values is called for. A cursor is typically marked bind sensitive, because there is a histogram
on the column with the bind variable. Since the presence of the histogram indicates that the
column is skewed, different values of the bind variable may call for different plans.
If a different bind value is used in a subsequent execution, it will use the same execution plan
because Oracle initially assumes it can be shared. However, the execution statistics for this new
bind value will be recorded and compared to the execution statistics for the previous value. If
Oracle determines that the new bind value caused the data volumes manipulated by the query to
be significantly different it "adapts" its behavior so that the same plan is not always shared for
this query. Hence a new plan will be generated based on the new bind value and the cursor is
marked bind-aware.
A bind-aware cursor may use different plans for different bind values, depending on how
selective the predicates containing the bind variable are. Note that the original cursor generated
for the statement will be discarded when the cursor switches to bind aware mode. This is a one-
time overhead. The cursor is marked as not shareable (is_shareable in V$SQL is "N"), which
means that this cursor will be among the first to be aged out of the cursor cache, and that it will
no longer be used.
When another new bind value is used, the optimizer tries to find a cursor it thinks will be a good
fit, based on similarity in the bind value's selectivity. If it cannot find such a cursor, it will create a
new one. If the plan for the new cursor is the same as an existing cursor, the two cursors will be
merged to save space in the cursor cache. And the selectivity range for that cursor will be
increased to include the selectivity of the new bind. This will result in one cursor being left

14

behind that is in a not shareable state. This cursor will again be aged out if there is crowding in
the cursor cache, and will not be used for future executions.

If you ever need to contact Oracle Support about a SQL issue obtaining a reproducible test case
is the single most important factor to ensure a speedy resolution. This can also be the longest and
most painful step. A new tool called the SQL Test Case Builder has been introduced in Oracle
Database 11g to help customers to gather as much information as possible relating to a SQL
incident and package it up ready to send to Oracle. This package of information will allow a
developer at Oracle to reproduce the problem standalone on a different Oracle instance and
resolve the issue sooner. You can access the SQL Test Case Builder through Oracle Enterprise
Manager or the PL/SQL package DBMS_SQLDIAG. There are two procedures relating to SQL Test
Case Builder; DBMS_SQLDIAG.EXPORT_SQL_TESTCASE, which enables you to export a SQL test
case for a given SQL statement into a given directory and
DBMS_SQLDIAG.IMPORT_SQL_TESTCASE, which enables you to import a given SQL test case
from a given directory.
To use SQL Test Case Builder, create an Oracle directory pointing to OS directory where the
output files will go.
SQL> CREATE DIRECTORY EXPDP AS '/scratch/mcolgan/spm/tc';

Then call the DBMS_SQLDIAG.EXPORT_SQL_TESTCASE package and pass it the directory you
just created and the SQL_ID for the statement in question. You will also need to supply a name
for the testcase; in this example we chose tc.

DECLARE tc clob;
BEGIN
DBMS_SQLDIAG.EXPORT_SQL_TESTCASE(
DIRECTORY=>'EXPDP',
SQL_ID=>'aqa1r0ca84rs1',
TESTCASE=>tc);
END;

After calling EXPORT_SQL_TESTCASE you will find several trace, dump and .sql files in the
specified directory, along with a README.txt. You should create a single compressed file from
this directory and upload it to Oracle Support. For security reason, the user data is not exported
by default. You have the option to set exportData to TRUE to include the data.

Oracle transforms SQL statements using a variety of sophisticated techniques during query
optimization. The purpose of this phase of query optimization is to transform the original SQL
statement into a semantically equivalent SQL statement that can be processed more efficiently.

15

In Oracle Database 11g, several new cost-based transformations were introduced. We discuss
below three of these transformations.

Group-by placement allows the optimizer to rewrite queries in order to minimize the number of
rows necessary for subsequent joins by performing the group-by operation before some of the
joins. For example, consider a query shown below.
SELECT p.prod_id, sum(s.quantity_sold)
FROM Products p, Sales s
WHERE p.prod_id = s.prod_id
GROUP BY p.prod_id;

In Oracle Database 10g the optimizer would have chosen a traditional plan of a hash join
followed by a group by for this query.

In Oracle Database 11g the Optimizer can transform this query to perform the group-by
operation before the join. The statement will be rewritten as follows:
SELECT V.sumv, p.prod_id
FROM Products p,
(SELECT sum(s.quantity_sold) as sumv, s.prod_id
FROM Sales s
GROUP BY s.prod_id) V
WHERE V.prod_id = p.prod_id

By rewriting the statement, the group by will now happens inside the view (VW_GBC_5 in the
plan below) thus reducing the number of rows that must be processed by the hash join between
SALES and PRODUCTS.

16

In previous releases when you had a SQL statement where a view V and a table T were joined by
a join predicate T.x = V.y, the Optimizer had only two possible join methods, a hash join or a
sort merge join to join T and V.
In Oracle Database 10g, we introduced the join predicate push down transformation, which
enabled the optimizer to push the join predicate into the view. So the join T.x = V.y becomes
T.x = T2.y (where T2 is the table inside view V, which has the column y in it) thereby opening
up the possibility of using a nested-loops join if an index is present on T2.y.
In Oracle Database 11g the join predicate push down capabilities have been extended to include
group by, distinct, anti-join, and semi-joins. For the following query;
SELECT p.prod_id, v1.row_count
FROM products p,
(SELECT s.prod_id, count(*) row_count
FROM sales s
WHERE s.quantity_sold BETWEEN 1 AND 47
GROUP BY s.prod_id) v1
WHERE p.supplier_id = 12
AND p.prod_id = v1.prod_id(+);

Although join predicate push down exists in Oracle Database 10g we can not use it due to the
group by. In the 10g we see the view v1 being evaluated followed by a hash join to the
PRODUCTS tables.

However, in 11g join predicate pushdown has become possible and we are now taking advantage
of the index on the sales table to do a nested-loops join instead of a hash join. The cost of the
new plan has come down from 79 to 28 because of join predicate pushdown.

17

A SQL statements that contains a NOT IN or NOT EXISTS subquery can often be rewritten
into a query containing an anti-join. An anti-join produces matches if the key from table on the
left hand side has no matches in the table on right hand side. Unnesting such a subquery can
provide an order of magnitude performance improvement. For example
SELECT C.CUST_ID, C.CUST_FIRST_NAME, C.CUST_LAST_NAME
FROM CUSTOMERS C
WHERE C.CUST_ID NOT IN (SELECT ST.CUST_ID
FROM SALES_TRANSACTIONS_EXT ST
WHERE ST.UNIT_PRICE < 15);
If the columns C.CUST_ID or ST.CUST_ID are nullable (allows a null value), then the
Optimizer cannot unnest the subquery in Oracle Database 10g using the existing (regular) anti-
join. Therefore, the execution plan effectively becomes a Cartesian product.

In Oracle Database 11g, a new variant of anti-join, called null-aware anti-join, has been
introduced; this allows the Optimizer to unnest the subquery, which means several new join
methods are possible including nested-loops join, a hash join or a sort-merge join. In this
example the 11g plan uses a hash join but in the execution plan it appears as a HASH JOIN
RIGHT ANTI NA, where “NA‟ stands for null-aware. Note that the plan cost has come down
from 639 to 38.

Undertaking a database upgrade is a daunting task for any DBA. Once the database has been
successfully upgraded you run the gauntlet of possible database behavior changes. On the top of
every DBA‟s list of potential behavior changes are execution plan changes. In order to easily
detect these changes, and rectify any execution plans that may have regressed, you need to have a
very good understanding of the execution plans and Optimizer statistics you had before you
began the upgrade. You also need to test your applications against the new release before
upgrading your production system.

18

With the introduction of SQL Plan Management (SPM) in 11g you have an additional safety net
to ensure execution plans do not change after the upgrade. In order to take full advantage of this
safety net, you need to capture your existing execution plans before you upgrade so they can be
used to seed SPM.

If you have access to SQL Tuning Sets (STS) in the diagnostics pack then this is the easiest way
to capture your existing 10g execution plans. An STS is a database object that includes one or
more SQL statements along with their execution statistics, execution context and their current
execution plan. An STS in Oracle Database 10gR1 will not capture the execution plans for the
SQL statements so it can‟t be used to seed SPM. Only a 10gR2 STS will capture the plans.
To begin you will need to create a new STS. You can either do this through Oracle Enterprise
Manager (EM) or using the DBMS_SQLTUNE package. In this example we will use DBMS_SQLTUNE
BEGIN
SYS.DBMS_SQLTUNE.CREATE_SQLSET (
sqlset_name => 'SPM_STS',
description => '10g plans');
END;

Once the STS has been created you need to populate it. You can populate an STS from the
workload repository, another STS, or from the cursor cache. In this case we will capture the SQL
statements and their execution plans from the cursor cache. This is a two-step process. In the
first step we create a ref cursor to select the specified SQL from the cursor cache (in this case all
non sys SQL statements). Then we use that ref cursor to populate the STS.
DECLARE
stscur dbms_sqltune.sqlset_cursor;
BEGIN
OPEN stscur FOR
SELECT VALUE(P)
FROM TABLE(dbms_sqltune.select_cursor_cache(
‘parsing_schema_name <> ‘‘SYS’’’,
null, null, null, null, 1, null, 'ALL')) P;

-- populate the sqlset
dbms_sqltune.load_sqlset(sqlset_name => 'SPM_STS',
populate_cursor => stscur);
END;

If you don‟t have access to SQL Tuning Sets you can capture your existing execution plan using
Stored Outlines. There are two ways to capture Stored Outlines, you can either manually create
one for each SQL statement using the CREATE OUTLINE command or let Oracle automatically

19

create a Stored Outline for each SQL statement that is executed. Below are the steps needed to
let Oracle automatically create the Stored Outlines for you.
1. Start a new session and issue the following command to switch on the automatic
capture of a Stored Outline for each SQL statement that gets parsed from now on until
you explicitly turn it off.
SQL > ALTER SYSTEM set CREATE_STORED_OUTLINES=OLDPLAN;

NOTE: Ensure that the user for which the Stored Outlines are to be created has the
CREATE ANY OUTLINE privilege. If they don‟t the Stored Outlines will not be captured.

2. Now execute your workload either by running your application or manually issuing SQL
statements. NOTE: if you manually issue the SQL statements ensure you use the exact
SQL text used by the application, if it uses bind variables you will have to use them too.
3. Once you have executed your critical SQL statements you should turn off the automatic
capture by issuing the following command:
SQL > ALTER SYSTEM set CREATE_STORED_OUTLINES=false;

4. To confirm you have captured the necessary Stored Outlines issue the following SQL
statement.
SQL> SELECT name, sql_text, category FROM user_outlines;
NOTE: Each Stored Outline should have the OLDPLAN category.
5. The actual Stored Outlines are stored in the OUTLN schema. Before you upgrade you
should export this schema as a backup.
exp outln/outln file=soutline.dmp owner=outln rows=y

Note: If you are not planning on doing an in-place upgrade you will have to move the STS or
Stored Outlines to the Oracle Database 11g system.

Before doing the upgrade you need to capture the current set of optimizer statistics. You will use
this set of statistics after the upgrade until the system is stable. You want to change as little as
possible during the upgrade in order to make it easier to diagnose any changes that may occur.
Since statistics have the biggest impact on the optimizer it is advisable for them to stay constant
during the upgrade. The best way to keep a backup of the statistics is to export a complete set of
Optimizer statistics into a statistics table using DBMS_STATS.EXPORT_*_STATS.
Begin by creating the stats table.
BEGIN
DBMS_STATS.CREATE_STATS_TABLE('SYS','MY_STATS_TAB');

20

END;
/

Then export the statistics for your critical schemas and select a stats_id to make it easy to identify
your 10g statistics.
BEGIN
DBMS_STATS.EXPORT_SCHEMA_STATS(‘SH’,'MY_STATS_TAB',’10g_stats’);
END;
/

Finally as an extra precaution, export the stats table as a backup.

It is absolutely vital that you test your complete application on Oracle Database 11g before you
upgrade. Testing will give you an insight into what might change after the upgrade and provide
you an opportunity to test your safety net (SPM). There is always the potential you will miss a
critical SQL statement when you gather your 10g plans, doing a full application test will give you
a chance to confirm you captured everything you need.
If your application currently uses optimizer hints then it is advisable to test the application, in
Oracle Database 11g, without the hint. Typically hints are added to a statement or an application
to work around a limitation or problem encountered in an earlier version of the Optimizer.
Often we find removing the hints gives better execution plans in newer releases of the database.
The easiest way to test without hints is to set the under score parameter _OPTIMIZER_IGNORE_
HINTS to TRUE.

Before you upgrade your production system to Oracle Database 11g you must collect and save
the following pieces of information to ensure you have a clear baseline that can be used for
future comparison if the need arises.
1. Gather Instance-wide performance statistics from the Production database (during peak load
times). These Instance-wide performance statistics include:
a. Statspack or AWR data and reports. You will need hourly reports for at least 7 days.
By default, AWR will take hourly snapshots that automatically capture the execution
plans for the top SQL statement and these reports will be retained for the last 8
days. However if you have changed the STATISTICS_LEVEL parameter or the
retention level you may need to change them back to default at least a week before
the upgrade. If you are using Statspack, you will need to configure it to take level 7
snapshots so you can collect segment statistics and plan information.
b. OS statistics including CPU, memory and IO (such as sar, vmstat, iostat)

21

2. Be sure to perform all business critical transactions as well as month-end processes and
common ad-hoc queries during the baseline capture.
3. Export the Statspack schema owner, PERFSTAT and keep the export file as backup. Note
this step is not necessary if you are using AWR.
4. Export a complete set of Optimizer statistics into a statistics table, and export the table as a
backup.
5. Make a backup of your init.ora file.
6. Capture the execution plan for key statements include the current Top SQL statements, and
any important SQL statements in the application. (See the capture existing execution plan
section above for details)

Once you have completed the software upgrade, but before you restart the applications and allow
users back on the system, you should populate SQL Plan Management (SPM) with the 10g
execution plans you captured before the upgrade. Seeding SPM with the 10g execution plans
ensures that the application will continue to use the same execution plans you had before the
upgrade. Any new execution plans found in Oracle Database 11g will be recorded in the plan
history for that statement but they will not be used. When you are ready you can evolve or verify
the new plans and only implement those that perform better than the 10g plan.

You can bulk load execution plans into SPM using four different techniques:
1. Populate the execution plans for a given SQL Tuning Set (STS)
2. Populate the execution plans from Stored Outlines
3. Use the execution plans currently in the Cursor Cache
4. Unpack existing SQL plan baselines from a staging table
Regardless of where you bulk loaded the plans from, each plan will be automatically accepted to
create a new SQL plan baseline or it will be added to an existing one. During an upgrade it‟s
most likely you will be bulk loading plan using the first three options. We will discuss each of
these options in detail below. More information on unpacking SQL plan baselines from a staging
table can be found in Chapter 15 of the Oracle Database Performance and Tuning Guide.

22

Execution plans can be bulk loaded from an STS into SPM using the PL/SQL procedure
DBMS_SPM.LOAD_PLANS_FROM_SQLSET or through Oracle Enterprise Manager (EM).
SQL> Variable cnt number
SQL> execute :cnt := DBMS_SPM.LOAD_PLANS_FROM_SQLSET( -
sqlset_name => 'SPM_STS');

In Oracle Database 11gR2 you can migrate stored outlines for one or more SQL statements to
SQL plan baselines using DBMS_SPM.MIGRATE_STORED_OUTLINE or through Oracle
Enterprise Manager (EM). You can specify which stored outline(s) to be migrated based on
outline name, SQL text, or outline category, or migrate all stored outlines in the system to SQL
plan baselines.
SQL> variable report clob;
-- Migrate a single Stored Outline by name
SQL> exec :report:=DBMS_SPM.MIGRATE_STORED_OUTLINE(
attribute_name=>'OUTLINE_NAME',
attribute_value => 'stmt01');

-- Migrate all Stored Outlines
SQL> exec :report:=DBMS_SPM.MIGRATE_STORED_OUTLINE(
attribute_name=>'ALL');

23

It is also possible to load plans for statements directly from the cursor cache into SPM. By
applying a filter - on the module name, the schema, or the SQL_ID - you can identify the SQL
statement or set of SQL statement you wish to capture. The plans can be loaded using the
PL/SQL procedure DBMS_SPM.LOAD_PLANS_FROM_CURSOR_CACHE or through Oracle
Enterprise Manager. Loading plans directly from the cursor cache can be extremely useful if you
were unable to capture plans for some or all of your statements prior to the upgrade.
By setting the parameter OPTIMIZER_FEATURES_ENABLE to the 10g version used before the
upgrade, you should be able to revert back to the same execution plans you had prior to the
upgrade. By capturing these 10g execution plans from the cursor cache you will be able to seed
SPM with the 10g plans before setting OPTIMIZER_FEATURES_ENABLE to your 11g version.
Note you must use the same Optimizer statistics you were using in 10g. Statistics should not be
re-gathered until all of the 10g plans have been captured.

Once SPM has been seeded with the pre-upgrade plans, you can check that the new create SQL
plan baselines are being used by looking at the notes section of the execution plan. If a SQL plan

24

baseline is in use you will see the phrase “SQL plan baseline XXXXX used for this statement” in
the note section.
SQL> explain plan for
2 SELECT *
3 FROM sh.sales
4 WHERE quantity_sold > 40
5 ORDER BY prod_id;

Explained.
SQL> SELECT * FROM table(dbms_xplan.display(null, null, 'basic +note'));

PLAN_TABLE_OUTPUT
-----------------------------------------------------------------
Plan hash value: 1421641795
--------------------------------------
| Id | Operation | Name |
--------------------------------------
| 0 | SELECT STATEMENT | |
| 1 | SORT ORDER BY | |
| 2 | PARTITION RANGE ALL| |
| 3 | TABLE ACCESS FULL | SALES |
--------------------------------------
Note
------
- SQL plan baseline "SQL_PLAN_2kgpw0an1uph654bc8843" used for this
statement

After the upgrade you want to use the 10g statistics until the system is stable. After the upgrade
and before you gather statistics for the first time you should switch on incremental statistics for
all partitioned tables. This will allow you to generate global statistics for partitioned tables, from
partition level statistics, and greatly reduce the amount of time it takes to gather statistics on
larger tables.
BEGIN
DBMS_STATS.SET_GLOBAL_PREFS('INCREMENTAL','TRUE');
END;

As a precautionary step, you may want to temporarily switch on pending statistics for all objects.
This will allow you to have an opportunity to test the new statistics before they are published and
start to be used in your production environment.
BEGIN
DBMS_STATS.SET_GLOBAL_PREFS(‘PENDING’,’TRUE’);
END;

At this point it is safe to gather 11g statistics. By not specifying any parameter values, the default
values will be use, which means it will automatically use the new statistics algorithm. All of the

25

statistics gathered will be stored as pending, thus allowing you to test them without impacting the
current environment.
BEGIN
DBMS_STATS.GATHER_SCHEMA_STATS(‘SH’);
END;

Alternatively you can export pending statistics from the production system and import them into
a test system, then test their impact before publishing them. Pending statistics can be exported
using procedure EXPORT_PENDING_STATS.
BEGIN
DBMS_STATS.EXPORT_PENDING_STATS(…);
END

You can now test your critical SQL statement with the pending statistics by changing the value of
the OPTIMIZER_USE_PENDING_STATISTICS parameter to TRUE at a session level. The
Optimizer will then use pending statistics for all SQL statements issued in this session.
ALTER SESSION set optimizer_use_pending_statistics=TRUE;

Finally when you have validated the 11g statistics you should turn off pending statistics and
publish the new statistics.
BEGIN
DBMS_STATS.SET_GLOBAL_PREFS(‘PENDING’,’FALSE’);
END

BEGIN
DBMS_STATS.PUBLISH_PENDING_STATS();
END

Once you have successfully upgraded to Oracle Database 11g and your application is up and
running again, you will have to monitor your environment carefully to ensure you do not
encounter any performance issues or plan regressions. The steps below outline what you should
do.
1. If you have not licensed the Diagnostic Pack, install or upgrade Statspack and set the
level to 7. Follow Statspack instructions specially if upgrading it.
2. Schedule Statspack snapshots every hour. This will let Statspack capture expensive SQL
in your 11g environment. If you have licensed the Diagnostic Pack then you can use the
AWR reports, which will be automatically captured hourly.
3. Capture OS statistics, which coincide with your Statspack reports.

26

4. Identify the expensive SQL (top SQL) using Statspack or AWR reports. Compare these
SQL statements to the top SQL statements you had prior to the upgrade. If they are not
the same you will need to investigate why.

a. If the pre-upgrade instance is still available, execute the source transaction in
both instances (10g and 11g ) and compare the execution plans, buffer gets,
CPU time, and total elapse times.

b. If pre-upgrade instance is no longer available, use the export of the
PERFSTAT schema (you took as part of the pre-upgrade check list) to find the
SQL statement and its execution plan (use script sprepsql.sql)

5. Determine root cause of sub optimal plans and take corrective action. Corrective action
may be in the form of: re-gathering statistics with different parameter settings, use SQL
Tuning Advisor, index creation, creation of a SQL Profile, use of Optimizer hints,
research of known Bugs, logging an SR, etc.

There is always a chance that you missed a SQL statement when you were recording the 10g
execution plans or a new module gets rolled out and you discover some poorly performing SQL
statement. In this cause you will need to correct the regressed SQL statement. If you don‟t have a
known good plan for this statement you will have to do some investigation into the cause of the
problem. In this section we will discuss the different options you have to correct this regressed
SQL statement.

When the execution plan for a SQL statement is sub-optimal, one of the first places you should
look is at the statistics used for that statement. The statistics have the biggest influence on the
optimizer and are the one aspect that will change over time as the data set changes.
The first thing you should check is that the statistics are accurate and up to date for every object
used in the query. Since Oracle Database 10g, the parameter OPTIMIZER_MODE is set to
ALL_ROWS by default. That means the cost-based Optimizer will always be used and any object
that does not have statistics will have statistics gathered at the time of hard parse using dynamic
sampling. Dynamic sampling is a great way to augment existing statistics but it isn‟t a good
substitute for genuine statistics. If dynamic sampling is being used, it will be shown in the note
section of the plan.

27

If the statistics are both current and accurate, then its possible that the sub-optimal plan is being
caused by real-world relationship in the data or by the presence of a function on one of the
columns in the where clause. You can address both of these situations by using the new extended
statistics feature in 11g.

You can invoke the SQL tuning advisor for a poor performing SQL statement, which will trigger
the Optimizer to analyze the problem SQL statement and recommend a solution. The tuning
process that is under taken is fully cost-based; it takes into account the past execution statistics of
the SQL statement and customizes the optimizer settings for that statement. It also collects
auxiliary information in conjunction with the regular statistics. The recommendation can fall into
one of the following categories;
Statistics Analysis: Check each object in the query for missing or stale statistics, and makes
recommendation to gather relevant statistics if they are needed.
SQL Profiling: The initial cardinality estimates used by the optimizer are checked and auxiliary
information is collected to remove all estimation errors. If the estimates are founded to be
lacking or if the auxiliary information proves to produce a more preformant execution plan a
SQL Profile using the auxiliary information will be generated and the end-user will be
recommend to create it.
Access Path Analysis: The Optimizer explores whether a new index can be used to significantly
improve access to each table in the query, and when appropriate makes recommendations to
create such indexes.
SQL Structure Analysis: The Optimizer tries to identify SQL statements that lend themselves
to bad plans, and makes relevant suggestions to restructure them.

If the statement you are investigating is failing with a critical error (ORA-600), then you run the
SQL Repair Advisor from the Problem Details page in Oracle Enterprise manger or by using
DBMS_SQLDIAG.CREATE_DIAGNOSIS_TASKS procedure. The SQL Repair Advisor will
investigate the SQL statement to see if it can come up with an alternative execution plan that will
not cause the critical error. If it finds such a plan it will recommend you apply a SQL patch for
this statement. Accepting the patch will allow your SQL statement to execute without hitting the
critical error but it may not be the most performant execution plan. Therefore you should still
open a Service Request with Oracle support and remove the SQL patch once you get an official
fix from Oracle.

28

In a small number of cases it may become necessary to correct a regressed SQL statement by
adding hints. However, for a third-party application it may not be possible to add hints to
statement and if you enable automatic plan capture in SQL Plan Management (SPM) you will
actually creating a new SQL plan baseline for the modified (hinted) statement and not for the
original statement. What you most likely want is to add this hinted plan to the plan history of the
original SQL statement. By using the simple steps below you can use SPM to capture the hinted
execution plan and associate it with the non-hinted SQL statement.
You begin by capturing a SQL plan baseline for the non-hinted SQL statement.
1. In a SQL*Plus session run the non-hinted SQL statement so we can begin the SQL plan
baseline capture
SQL> SELECT prod_name, SUM(amount_sold)
FROM Sales s, Products p
WHERE s.prod_id=p.prod_id
AND prod_category = :ctgy
GROUP BY prod_name;

2. Then find the SQL_ID for the statement in the V$SQL view.

SQL> SELECT sql_id, sql_fulltext
FROM V$SQL

29

WHERE sql_text LIKE '%SELECT prod_name, SUM(%';

SQL_ID SQL_FULLTEXT
------------- ---------------------------------------
74hnd835n81yv select SQL_ID, SQL_FULLTEXT from v$SQL
chj6q8z7ykbyy SELECT PROD_NAME, SUM(AMOUNT_SOLD)

3. Using the SQL_ID create a SQL plan baseline for the statement.
SQL> variable cnt number;
SQL> EXECUTE :cnt :=DBMS_SPM.LOAD_PLAN_FROM_CURSOR_CACHE(
sql_id=>'chj6q8z7ykbyy');

4. The plan that was captured is the sub-optimal plan and it will need to be disabled. The
SQL_HANDLE & PLAN_NAME are required to disable the plan. These can found by looking in
DBA_SQL_PLAN_BASELINE view.

SQL> SELECT sql_handle, sql_text, plan_name, enabled
FROM dba_sql_plan_baselines;

SQL_HANDLE SQL_TEXT PLAN_NAME ENABLE
------------------------ ---------------------- ------------------------------ ---
SYS_SQL_bf5c9b08f72bde3e SELECT PROD_NAME,SUM SQL_PLAN_byr4v13vkrrjy42949306 YES

5. Using DBMS_SPM.ALTER_SQL_PLAN_BASELINE disable the bad plan

SQL> variable cnt number;
SQL> exec :cnt :=DBMS_SPM.ALTER_SQL_PLAN_BASELINE(
SQL_HANDLE => 'SYS_SQL_bf5c9b08f72bde3e',
PLAN_NAME => 'SQL_PLAN_byr4v13vkrrjy42949306',
ATTRIBUTE_NAME => 'enabled',
ATTRIBUTE_VALUE => 'NO');

SQL> SELECT sql_handle, sql_text, plan_name, enabled
FROM dba_sql_plan_baselines;

SQL_HANDLE SQL_TEXT PLAN_NAME ENABLE
------------------------ ---------------------- ------------------------------ ---
SYS_SQL_bf5c9b08f72bde3e SELECT PROD_NAME,SUM SQL_PLAN_byr4v13vkrrjy42949306 NO

6. Now you need to modify the SQL statement using the necessary hints & execute the modified
statement.

SQL> SELECT /*+ INDEX(p) */ prod_name, SUM(amount_sold)
FROM Sales s, Products p
WHERE s.prod_id=p.prod_id
AND prod_category = :ctgy

30

GROUP BY prod_name;

7. Find the SQL_ID and PLAN_HASH_VALUE for the hinted SQL statement in the V$SQL view.

SQL> SELECT sql_id, plan_hash_value, fulltext
FROM V$SQL
WHERE sql_text LIKE '%SELECT /*+ INDEX(p) */ prod_na%';

SQL_ID PLAN_HASH_VALUE SQL_FULLTEXT
------------- --------------- ----------------------------

9t5v8swp79svs 3262214722 select SQL_ID, SQL_FULLTEXT from v$SQL
djkqjd0kvgmb5 3074207202 SELECT /*+ INDEX(p) */ PROD_NAME,

8. Using the SQL_ID and PLAN_HASH_VALUE for the modified plan, create a new accepted
plan for original SQL statement by associating the modified plan to the original statement's
SQL_HANDLE.

exec :cnt:=dbms_spm.load_plans_from_cursor_cache(
sql_id => 'djkqjd0kvgmb5',
plan_hash_value => 3074207202,
sql_handle => 'SYS_SQL_bf5c9b08f72bde3e‘);

Since the introduction of the cost-based optimizer (CBO) in Oracle 7.0, people have been
fascinated by the CBO and the statistics that it feeds on. It has long been felt that the internals of
the CBO were shrouded in mystery and that a degree in wizardry was needed to work with it. In
Oracle Database 11g the CBO‟s effectiveness and ease of use have been greatly improved. Most
notably, the introduction of SQL Plan Management (SPM), which has simplified the daunting
task of upgrading the database by ensuring that the Optimizer will use only the execution plans
you had before the upgrade, if the new execution plans regress.
By outlining in detail the changes made in this release to the CBO and its statistics, we hope to
remove some of the mystery that surrounds them and help make the upgrade process smoother,
as being forewarned is being forearmed.

31

Twp Upgrading 10g To 11g What To Expect From Optimizer

Recommended

More Related Content

What's hot (20)

Similar to Twp Upgrading 10g To 11g What To Expect From Optimizer (20)

Twp Upgrading 10g To 11g What To Expect From Optimizer