chapter 4 an introduction to sql. copyright © 2004 pearson addison-wesley. all rights reserved.4-2...

Chapter 4

An Introduction to SQL

Copyright © 2004 Pearson Addison-Wesley. All rights reserved. 4-2

Topics in this Chapter

• SQL: History and Overview • The Catalog • Views• Embedded SQL• Cursors• Dynamic SQL• Call-Level Interface


History of SQL

• Developed by IBM in 1970’s • Early prototype called System R• Current version is SQL:1999• All major database vendors support SQL• All support a superset of a subset of SQL• Both relvar and relation = table in SQL• Tuple = row in SQL• Attribute = column in SQL


Overview of SQL

• It is the standard language for relational systems, although imperfect

• Supports data definition (DDL) and data manipulation (DML)

• DDL: CREATE TABLE, CREATE TYPE, ALTER TABLE

• DML: SELECT, INSERT, DELETE, UPDATE


Relational operators in SQL

• Restrict is implemented by SELECT– Subset of rows

– Uses WHERE clause to narrow result

– Uses SELECT * to copy entire table

• Project is also implemented by SELECT– Subset of columns


Relational operators in SQL

• Join is also implemented by SELECT– Result is a single table matched by values in the

join columns of each source table

– Uses WHERE clause to specify source tables

– Uses dot operator to qualify like column names


Update operators in SQL - Insert

• Insert is implemented by INSERT INTO– Ex,: INSERT INTO TEMP (P#, Weight)

SELECT P#, Weight FROM P

WHERE COLOR = COLOR (‘Red’);

– Inserts part number and weight from P into part number and weight in TEMP for all red parts


Update operators in SQL - Delete

• Delete is implemented by DELETE– Ex,: DELETE FROM SP

WHERE P# = P# (‘P2’);

– Deletes all shipments for part P2


Update operators in SQL - Update

• Update is implemented by UPDATE– Ex,: UPDATE S

SET STATUS = 2 * STATUS,

CITY = ‘Rome’

WHERE CITY = ‘Paris’;

– Doubles the status of the Parisian suppliers and moves them to Rome.


The Catalog in SQL

• Every database has a catalog• A catalog has many schemas, one per user• Each catalog includes one Information

Schema• The Information Schema is a collection of all

other schemas as views• Represents a hypothetical Definition Schema

for all users


The Information Schema - Examples

• Tables (includes views)• Views (does not include tables)• Table constraints• Referential constraints• Column constraints• Privileges• (Because these objects are themselves tables,

they can be queried)


Views

• CREATE VIEW GOOD_SUPPLIER

AS SELECT S#, STATUS, CITY

FROM S

WHERE STATUS > 15;• You can now operate on the view• SELECT S#, STATUS

FROM GOOD_SUPPLIER

WHERE CITY = ‘London’;


Views - Example

• Creating a view, and then selecting from the view is equivalent to a select sub-query

• SELECT GOOD_SUPPLIER .S#, GOOD_SUPPLIER .STATUS

FROM (SELECT S#, STATUS, CITYFROM S WHERE STATUS > 15 ) AS GOOD_SUPPLIER;

WHERE GOOD_SUPPLIER.CITY = ‘London’;


Embedded SQL

• SQL statements can be embedded in a host program

• C, COBOL, PL/1, Java, and many others• Initiated in host via EXEC SQL statement• Any SQL statement is valid after EXEC SQL


Embedded SQL

• SQL statements after EXEC SQL can reference host variables

• Declare any such variables in a SQL DECLARE section

• Then, prefix : (colon) to the variable name when you reference it


Embedded SQL - Example

EXEC SQL BEGIN DECLARE SECTION;

DCL SQLSTATE CHAR(5);

DCL P# CHAR(6);

DCL WEIGHT DECIMAL(5,1);

EXEC SQL END DECLARE SECTION;

P# = ‘P2’;

EXEC SQL SELECT P.WEIGHT

INTO :WEIGHT

FROM P

WHERE P.P# = (:P#);


Embedded SQL - SQLSTATE

IF SQLSTATE = ‘00000’

THEN …

ELSE …

• Every host language must include a host variable called SQLSTATE that returns 00000 is execution is ok

• If no data found then it returns 02000


Embedded SQL – Loosely coupled with the host

• Data types between host and tables must be compatible

• Host variables can have the same name as table columns, but need not

• SQL retrieval returns sets of rows• Host languages handle one row at a time• To allow this, SQL sets up a CURSOR to hold

returned row sets, so they can be processed one row at a time


Embedded SQL – Sometimes Cursors are not Needed

• SELECT, when it returns at most one row• INSERT, however many rows• DELETE, likewise• UPDATE, ditto• Example:

EXEC SQL UPDATE S

SET STATUS = STATUS + :RAISE

WHERE CITY = ‘London’;


Embedded SQL – Cursors

• A cursor is a buffer holding a row set• The buffer is called the “context area”• A cursor maintains a pointer to the currently

active row, or record, and begins before, yes, before the first row

• You can process each row, one at a time, and then advance the cursor

• UPDATE and DELETE are supported when you reference the “CURRENT” row


Embedded SQL –Cursors

• A cursor must be declared – here is where you place your SQL statement that will return a set

• Next OPEN the cursor • You may then FETCH the cursor into

variables you have declared• Each FETCH reads a row into your variables

– Often this is accomplished within a loop

• Your first FETCH reads the first row• CLOSE when finished, please


Dynamic SQL

• Embedded SQL can be constructed and invoked dynamically

DCL SQLSOURCE CHAR VARYING (65000);

SQLSOURCE = ‘DELETE FROM SP WHERE QTY < QTY (300)’;

EXEC SQL PREPARE SQLPREPPED FROM :SQLSOURCE;

EXEC SQL EXECUTE SQLPREPPED;


SQL Call-Level Interface (CLI)

• Alternative to Embedded SQL• Similar functionality, based in MS ODBC• Better than Embedded SQL for two reasons

– Can be compiled by the host compiler and distributed shrink wrapped to customers

– Greater DBMS independence because they can interface with anyone’s SQL engine

chapter 4 an introduction to sql. copyright © 2004 pearson addison-wesley. all rights reserved.4-2...

Documents