This article has the following sections:

• Part 1 - Using SQL CBO Parameters
• Part 2 - Using CBO statistics
• Part 3 - Proper SQL development environment
• Part 4 - Using histograms to tune SQL
• Part 5 - Clustering and SQL tuning
• Part 6 - External costs and SQL execution
• Part 7 - Using hints to tune SQL
• Part 8 - Locating sub-optimal SQL

PART 7 - Using Hints to Change Execution Plans

As the optimizer becomes more sophisticated with each release, Oracle provides an increasing number of methods for changing the execution plans for your SQL. The most common use for Oracle hints is as a debugging tool. You can use the hints to determine the optimal execution plan, and then work backward, adjusting the statistics to make the vanilla SQL simulate the hinted query.

Using Oracle hints can be very complicated and Oracle developers only use hints as a last resort, preferring to alter the statistics to change the execution plan. Oracle contains more than 124 hints, and many of them are not found in the Oracle documentation. (See Listing 2)

Listing 2: Documented Oracle Hints:

Undocumented Hints:

Let's take a quick look at how hints are used to alter optimizer execution plans: A optimizer hint is an optimizer directive placed inside comments inside your SQL statement and used in those rare cases where the optimizer makes an incorrect decision about the execution plan. Because hints are inside comments, it is important to ensure that the hint name is spelled correctly and that the hint is appropriate to the query.

The all_rows optimizer mode is designed to minimize computing resources and it favors full-table scans.  Index access (first_rows) adds additional I/O overhead, but they return rows faster, back to the originating query:

For example, the following hint is invalid because first_rows access and parallel access are mutually exclusive. That's because parallel always assumes a full-table scan and first_rows favors index access.

-- An invalid hint
select /*+ first_rows parallel(emp,8)*/ 
   emp_name 
from 
   emp 
order by 
   ename;

Some Oracle professionals will place hints together to reinforce their wishes. For example, if we have an SMP server with eight or more CPUs, we may want to use Oracle Parallel Query to speed-up legitimate full-table scans. When using parallel query, we seldom want to turn-on parallelism at the table level (alter table customer parallel 35;) because the setting of parallelism for a table influences the optimizer, causing the optimizer to see full-table scan is inexpensive. Hence, most Oracle professionals specify parallel query on a   query-by-query basis, combining the full hint with the parallel hint to ensure a fast parallel full-table scan:

-- A valid hint
select /*+ full parallel(emp,35)*/ 
   emp_name 
from 
   emp 
order by 
   ename;

Now that we have the general concept of hints, let's take a look at one of the most important hints for optimizer tuning.

The ordered hint determines the driving table for the query execution and also specifies the order that tables are joined together. The ordered hint requests that the tables should be joined in the order that they are specified in the from clause, with the first table in the from clause specifying the driving table. Using the ordered hint can save a huge amount of parse time and speed SQL execution because you are telling the optimizer the best order to join the tables.

For example, the following query uses the ordered hint to join the tables in their specified order in the from clause. In this example, we further refine the execution plan by specifying that the emp to dept join use a hash join and the sal to bonus join use a nested loop join:

select 
/*+ ordered use_hash (emp, dept) use_nl (sal, bon) */
from
   emp,
   dept,
   sal,
   bon
where . . .

Of course, the ordered hint is most commonly used in data warehouse queries or in SQL that joins more than five tables.

Next let's look at another last resort, the adjustment of Oracle parameters. Oracle does not recommend changing any of these parameters, except as a last resort. However, it is interesting to see how these parameters change the way that the optimizer determines execution plans.

Our preferred SQL tuning tool:

Our WISE tool is the easiest way to analyze Oracle SQL performance and WISE allows you to spot hidden SQL performance trends.

Need an Oracle Health Check? Do you have bad performance after an upgrade?   Need to certify that your database follows best practices? BC Oracle performance gurus can quickly certify every aspect of your Oracle database and provide a complete verification that your database is fully optimized.

Donald K. Burleson [info@remote-dba.net] is one of the world's most widely-read Oracle database experts. He has written 19 books, published more than 100 articles in national magazines, and serves as editor-in-chief of Oracle Internals, a leading Oracle database journal. Burleson's latest book is Creating a Self-Tuning Database by Rampant TechPress. Don's Web sites are http://www.dba-oracle.com , http://www.remote-dba.net/ and
http://rampant.cc .