1 Overview of PL/SQL

Tight Integration with SQL

PL/SQL is tightly integrated with SQL, the most widely used database manipulation language. For example:

• PL/SQL lets you use all SQL data manipulation, cursor control, and transaction control statements, and all SQL functions, operators, and pseudocolumns.

• PL/SQL fully supports SQL data types.

  You need not convert between PL/SQL and SQL data types. For example, if your PL/SQL program retrieves a value from a column of the SQL type VARCHAR2, it can store that value in a PL/SQL variable of the type VARCHAR2.

  You can give a PL/SQL data item the data type of a column or row of a database table without explicitly specifying that data type (see "%TYPE Attribute" and "%ROWTYPE Attribute").

• PL/SQL lets you run a SQL query and process the rows of the result set one at a time (see "Processing a Query Result Set One Row at a Time").

• PL/SQL functions can be declared and defined in the WITH clauses of SQL SELECT statements (see Oracle Database SQL Language Reference).

PL/SQL supports both static and dynamic SQL. Static SQL is SQL whose full text is known at compile time. Dynamic SQL is SQL whose full text is not known until run time. Dynamic SQL lets you make your applications more flexible and versatile. For more information, see Chapter 6, "PL/SQL Static SQL" and Chapter 7, "PL/SQL Dynamic SQL".

High Performance

PL/SQL lets you send a block of statements to the database, significantly reducing traffic between the application and the database.

Bind Variables

When you embed a SQL INSERT, UPDATE, DELETE, MERGE, or SELECT statement directly in your PL/SQL code, the PL/SQL compiler turns the variables in the WHERE and VALUES clauses into bind variables (for details, see "Resolution of Names in Static SQL Statements"). Oracle Database can reuse these SQL statements each time the same code runs, which improves performance.

PL/SQL does not create bind variables automatically when you use dynamic SQL, but you can use them with dynamic SQL by specifying them explicitly (for details, see "EXECUTE IMMEDIATE Statement").

Subprograms

PL/SQL subprograms are stored in executable form, which can be invoked repeatedly. Because stored subprograms run in the database server, a single invocation over the network can start a large job. This division of work reduces network traffic and improves response times. Stored subprograms are cached and shared among users, which lowers memory requirements and invocation overhead. For more information about subprograms, see "Subprograms".

Optimizer

The PL/SQL compiler has an optimizer that can rearrange code for better performance. For more information about the optimizer, see "PL/SQL Optimizer".

High Productivity

PL/SQL lets you write compact code for manipulating data. Just as a scripting language like PERL can read, transform, and write data in files, PL/SQL can query, transform, and update data in a database.

PL/SQL has many features that save designing and debugging time, and it is the same in all environments. If you learn to use PL/SQL with one Oracle tool, you can transfer your knowledge to other Oracle tools. For example, you can create a PL/SQL block in SQL Developer and then use it in an Oracle Forms trigger. For an overview of PL/SQL features, see "Main Features of PL/SQL".

Portability

You can run PL/SQL applications on any operating system and platform where Oracle Database runs.

Scalability

PL/SQL stored subprograms increase scalability by centralizing application processing on the database server. The shared memory facilities of the shared server let Oracle Database support thousands of concurrent users on a single node. For more information about subprograms, see "Subprograms".

For further scalability, you can use Oracle Connection Manager to multiplex network connections. For information about Oracle Connection Manager, see Oracle Database Net Services Reference.

Manageability

PL/SQL stored subprograms increase manageability because you can maintain only one copy of a subprogram, on the database server, rather than one copy on each client system. Any number of applications can use the subprograms, and you can change the subprograms without affecting the applications that invoke them. For more information about subprograms, see "Subprograms".

Support for Object-Oriented Programming

PL/SQL supports object-oriented programming with "Abstract Data Types".

Error Handling

PL/SQL makes it easy to detect and handle errors. When an error occurs, PL/SQL raises an exception. Normal execution stops and control transfers to the exception-handling part of the PL/SQL block. You do not have to check every operation to ensure that it succeeded, as in a C program. For more information, see Chapter 11, "PL/SQL Error Handling".

Blocks

The basic unit of a PL/SQL source program is the block, which groups related declarations and statements.

A PL/SQL block is defined by the keywords DECLARE, BEGIN, EXCEPTION, and END. These keywords divide the block into a declarative part, an executable part, and an exception-handling part. Only the executable part is required. A block can have a label.

Example 1-1 shows the basic structure of a PL/SQL block. For syntax details, see "Block".

<< label >> (optional)
DECLARE    -- Declarative part (optional)
  -- Declarations of local types, variables, & subprograms

BEGIN      -- Executable part (required)
  -- Statements (which can use items declared in declarative part)

[EXCEPTION -- Exception-handling part (optional)
  -- Exception handlers for exceptions (errors) raised in executable part]
END;

Declarations are local to the block and cease to exist when the block completes execution, helping to avoid cluttered namespaces for variables and subprograms.

Blocks can be nested: Because a block is an executable statement, it can appear in another block wherever an executable statement is allowed.

You can submit a block to an interactive tool (such as SQL*Plus or Enterprise Manager) or embed it in an Oracle Precompiler or OCI program. The interactive tool or program runs the block one time. The block is not stored in the database, and for that reason, it is called an anonymous block (even if it has a label).

An anonymous block is compiled each time it is loaded into memory, and its compilation has three stages:

1. Syntax checking: PL/SQL syntax is checked, and a parse tree is generated.

2. Semantic checking: Type checking and further processing on the parse tree.

3. Code generation

Variables and Constants

PL/SQL lets you declare variables and constants, and then use them wherever you can use an expression. As the program runs, the values of variables can change, but the values of constants cannot. For more information, see "Declarations" and "Assigning Values to Variables".

Subprograms

A PL/SQL subprogram is a named PL/SQL block that can be invoked repeatedly. If the subprogram has parameters, their values can differ for each invocation. PL/SQL has two types of subprograms, procedures and functions. A function returns a result. For more information about PL/SQL subprograms, see Chapter 8, "PL/SQL Subprograms."

PL/SQL also lets you invoke external programs written in other languages. For more information, see "External Subprograms".

Packages

A package is a schema object that groups logically related PL/SQL types, variables, constants, subprograms, cursors, and exceptions. A package is compiled and stored in the database, where many applications can share its contents. You can think of a package as an application.

You can write your own packages—for details, see Chapter 10, "PL/SQL Packages." You can also use the many product-specific packages that Oracle Database supplies. For information about these, see Oracle Database PL/SQL Packages and Types Reference.

Triggers

A trigger is a named PL/SQL unit that is stored in the database and run in response to an event that occurs in the database. You can specify the event, whether the trigger fires before or after the event, and whether the trigger runs for each event or for each row affected by the event. For example, you can create a trigger that runs every time an INSERT statement affects the EMPLOYEES table.

For more information about triggers, see Chapter 9, "PL/SQL Triggers."

Input and Output

Most PL/SQL input and output (I/O) is done with SQL statements that store data in database tables or query those tables. For information about SQL statements, see Oracle Database SQL Language Reference.

All other PL/SQL I/O is done with PL/SQL packages that Oracle Database supplies, which Table 1-1 summarizes.

To display output passed to DBMS_OUTPUT, you need another program, such as SQL*Plus. To see DBMS_OUTPUT output with SQL*Plus, you must first issue the SQL*Plus command SET SERVEROUTPUT ON.

Some subprograms in the packages in Table 1-1 can both accept input and display output, but they cannot accept data directly from the keyboard. To accept data directly from the keyboard, use the SQL*Plus commands PROMPT and ACCEPT.

Data Abstraction

Data abstraction lets you work with the essential properties of data without being too involved with details. You can design a data structure first, and then design algorithms that manipulate it.

Topics

• Cursors

• Composite Variables

• %ROWTYPE Attribute

• %TYPE Attribute

• Abstract Data Types

Control Statements

Control statements are the most important PL/SQL extension to SQL.

PL/SQL has three categories of control statements:

• Conditional selection statements, which let you run different statements for different data values.

  For more information, see "Conditional Selection Statements".

• Loop statements, which let you repeat the same statements with a series of different data values.

  For more information, see "LOOP Statements".

• Sequential control statements, which allow you to go to a specified, labeled statement, or to do nothing.

  For more information, see "Sequential Control Statements".

Conditional Compilation

Conditional compilation lets you customize the functionality in a PL/SQL application without removing source text. For example, you can:

• Use new features with the latest database release, and disable them when running the application in an older database release.

• Activate debugging or tracing statements in the development environment, and hide them when running the application at a production site.

For more information, see "Conditional Compilation".

Processing a Query Result Set One Row at a Time

PL/SQL lets you issue a SQL query and process the rows of the result set one at a time. You can use a basic loop, as in Example 1-2, or you can control the process precisely by using individual statements to run the query, retrieve the results, and finish processing.

BEGIN
  FOR someone IN (
    SELECT * FROM employees
    WHERE employee_id < 120
    ORDER BY employee_id
  )
  LOOP
    DBMS_OUTPUT.PUT_LINE('First name = ' || someone.first_name ||
                         ', Last name = ' || someone.last_name);
  END LOOP;
END;
/

First name = Steven, Last name = King
First name = Neena, Last name = Kochhar
First name = Lex, Last name = De Haan
First name = Alexander, Last name = Hunold
First name = Bruce, Last name = Ernst
First name = David, Last name = Austin
First name = Valli, Last name = Pataballa
First name = Diana, Last name = Lorentz
First name = Nancy, Last name = Greenberg
First name = Daniel, Last name = Faviet
First name = John, Last name = Chen
First name = Ismael, Last name = Sciarra
First name = Jose Manuel, Last name = Urman
First name = Luis, Last name = Popp
First name = Den, Last name = Raphaely
First name = Alexander, Last name = Khoo
First name = Shelli, Last name = Baida
First name = Sigal, Last name = Tobias
First name = Guy, Last name = Himuro
First name = Karen, Last name = Colmenares

PL/SQL Engine

The PL/SQL compilation and runtime system is an engine that compiles and runs PL/SQL units. The engine can be installed in the database or in an application development tool, such as Oracle Forms.

In either environment, the PL/SQL engine accepts as input any valid PL/SQL unit. The engine runs procedural statements, but sends SQL statements to the SQL engine in the database, as shown in Figure 1-1.

Figure 1-1 PL/SQL Engine

Description of "Figure 1-1 PL/SQL Engine"

Typically, the database processes PL/SQL units.

When an application development tool processes PL/SQL units, it passes them to its local PL/SQL engine. If a PL/SQL unit contains no SQL statements, the local engine processes the entire PL/SQL unit. This is useful if the application development tool can benefit from conditional and iterative control.

For example, Oracle Forms applications frequently use SQL statements to test the values of field entries and do simple computations. By using PL/SQL instead of SQL, these applications can avoid calls to the database.

PL/SQL Units and Compilation Parameters

A PL/SQL unit is one of these:

• PL/SQL anonymous block

• FUNCTION

• LIBRARY

• PACKAGE

• PACKAGE BODY

• PROCEDURE

• TRIGGER

• TYPE

• TYPE BODY

PL/SQL units are affected by PL/SQL compilation parameters (a category of database initialization parameters). Different PL/SQL units—for example, a package specification and its body—can have different compilation parameter settings.

Table 1-2 summarizes the PL/SQL compilation parameters. To display the values of these parameters for specified or all PL/SQL units, query the static data dictionary view ALL_PLSQL_OBJECT_SETTINGS. For information about this view, see Oracle Database Reference.

The compile-time values of the parameters in Table 1-2 are stored with the metadata of each stored PL/SQL unit, which means that you can reuse those values when you explicitly recompile the unit. (A stored PL/SQL unit is created with one of the "CREATE [ OR REPLACE ] Statements". An anonymous block is not a stored PL/SQL unit.)

To explicitly recompile a stored PL/SQL unit and reuse its parameter values, you must use an ALTER statement with both the COMPILE clause and the REUSE SETTINGS clause. For more information about REUSE SETTINGS, see "compiler_parameters_clause". (All ALTER statements have this clause. For a list of ALTER statements, see "ALTER Statements".)