Parsing the SQL Statement Oracle Tips by Burleson Consulting

When a request is made by a program to retrieve data columns, Oracle receives the SQL statement and places it into the library cache areas of the shared pool. Once the statement is in the shared pool, Oracle invokes the parser to validate the SQL syntax. The basic purpose of the SQL parse is to validate the structure of the SQL syntax, verify that the user is authorized to view the data, and in some cases, reformulate the query to make it more efficient for the optimizer.

The process of parsing the SQL statement is illustrated in Figure 3-1.

Figure 1: The parsing of a SQL statement

While the basic process of parsing SQL is very straightforward, there are several optimal processes that will rewrite a SQL statement to make a faster execution plan.

• Query_rewrite_enabled This initialization parameter directs Oracle to rewrite data warehouse queries to prevent resummarization of large volumes of data.

• Cursor_sharing When set to FORCE, this initialization parameter directs Oracle to replace literal values in SQL statements with host variables.

Oracle has recognized that SQL statements can sometimes be restructured to allow the optimizer to generate a more favorable execution plan. Let’s take a closer look at these rewrite features.

Query Rewrite

Oracle has a special feature called materialized views that can greatly speed-up data warehouse queries. In a materialized view, a summary table is created from a base table, and all queries that perform a similar summation against the base table will be transparently rewritten to reference the pre-built summary table.

What follows is a simple example. We begin by creating a materialized view that sums sales data.

create materialized view
   sum_sales
build immediate
refresh complete
enable query rewrite
as
select
   product_nbr,
   sum(sales) sum_sales
from
   sales;

Now, when we have any query that summarizes sales, that query will be dynamically rewritten to reference the summary table.

alter session set query_rewrite_enabled=true;
set autotrace on

select
   sum(sales)
from  
   sales
;

In the execution plan for this query we see that the sum_sales table is being referenced.

Execution Plan
----------------------------------------------------------
   0      SELECT STATEMENT Optimizer=CHOOSE (Cost=1 Card=1 Bytes=83)
   1    0   SORT (AGGREGATE)
   2    1     TABLE ACCESS (FULL) OF 'SUM_SALES' (Cost=1 Card=423 Bytes=5342)

Note: If you use bind variables in a query, the query will be not be rewritten to use materialized views even if you have enabled query rewrite.

Once the query rewrite feature is enabled, you can use standard SQL hints to force the SQL parser to rewrite the query.

select /*+REWRITE(sales)*/
...

As Oracle SQL evolves and becomes more sophisticated, there will be more cases where the SQL parser will rewrite queries into a more efficient form. Next, let’s look at a method for making literal SQL reusable inside the library cache.

This is an excerpt from "Oracle High-Performance SQL Tuning" by Donald K. Burleson, published by Oracle Press.