RNDr. Michal Kopecký, Ph.D. Department of Software Engineering, Faculty of Mathematics and Physics, Charles University in Prague

Similar documents
RNDr. Michal Kopecký, Ph.D. Department of Software Engineering, Faculty of Mathematics and Physics, Charles University in Prague

Active Databases Part 1: Introduction CS561

Trigger is a stored procedure which is called implicitly by oracle engine whenever a insert, update or delete statement is fired.

5 Integrity Constraints and Triggers

RNDr. Michal Kopecký, Ph.D. Department of Software Engineering, Faculty of Mathematics and Physics, Charles University in Prague

doc. RNDr. Tomáš Skopal, Ph.D. RNDr. Michal Kopecký, Ph.D.

RDBMS Topic 4 Adv. SQL, MSBTE Questions and Answers ( 12 Marks)

Database Management Systems Triggers

NULLs & Outer Joins. Objectives of the Lecture :

COSC344 Database Theory and Applications. Lecture 11 Triggers

5. Single-row function

GIFT Department of Computing Science Data Selection and Filtering using the SELECT Statement

Creating and Managing Tables Schedule: Timing Topic

Databases. Relational Model, Algebra and operations. How do we model and manipulate complex data structures inside a computer system? Until

New York Oracle Users Group. September 26, 2002 New York, NY

PL/SQL. Exception. When the PL/SQL engine cannot execute the PLSQL block it raise an error. Every Oracle error has an error number

ORACLE VIEWS ORACLE VIEWS. Techgoeasy.com

Oracle Class VII More on Exception Sequence Triggers RowID & Rownum Views

Enhanced Data Models for Advanced Applications. Active Database Concepts

Objectives. After completing this lesson, you should be able to do the following:

doc. RNDr. Tomáš Skopal, Ph.D. doc. RNDr. Irena Mlýnková, Ph.D. RNDr. Michal Kopecký, Ph.D.

Databases - 4. Other relational operations and DDL. How to write RA expressions for dummies

Oracle Database 11g: SQL and PL/SQL Fundamentals

SQL Structured Query Language Introduction

Database Systems: Design, Implementation, and Management Tenth Edition. Chapter 7 Introduction to Structured Query Language (SQL)

Oracle Syllabus Course code-r10605 SQL

CS2 Current Technologies Lecture 3: SQL - Joins and Subqueries

Interview Questions on DBMS and SQL [Compiled by M V Kamal, Associate Professor, CSE Dept]

MANAGING DATA(BASES) USING SQL (NON-PROCEDURAL SQL, X401.9)

SQL Interview Questions

SYSTEM CODE COURSE NAME DESCRIPTION SEM

RNDr. Michal Kopecký, Ph.D. Department of Software Engineering, Faculty of Mathematics and Physics, Charles University in Prague

SQL. Rodrigo García Carmona Universidad San Pablo-CEU Escuela Politécnica Superior

CHAPTER: 4 ADVANCE SQL: SQL PERFORMANCE TUNING (12 Marks)

Lab # 6. Using Subqueries and Set Operators. Eng. Alaa O Shama

Chapter 7. Introduction to Structured Query Language (SQL) Database Systems: Design, Implementation, and Management, Seventh Edition, Rob and Coronel

MANAGING DATA(BASES) USING SQL (NON-PROCEDURAL SQL, X401.9)

Now, we can refer to a sequence without having to use any SELECT command as follows:

SQL functions fit into two broad categories: Data definition language Data manipulation language

Database Systems: Design, Implementation, and Management Tenth Edition. Chapter 8 Advanced SQL

Oracle Database: SQL and PL/SQL Fundamentals NEW

UNIT-IV (Relational Database Language, PL/SQL)

Relational Model History. COSC 416 NoSQL Databases. Relational Model (Review) Relation Example. Relational Model Definitions. Relational Integrity

Oracle Database: SQL and PL/SQL Fundamentals Ed 2

Programming the Database

Lab # 2. Data Definition Language (DDL) Eng. Alaa O Shama

Overview. Data Integrity. Three basic types of data integrity. Integrity implementation and enforcement. Database constraints Transaction Trigger

1 Prepared By Heena Patel (Asst. Prof)

Chapter 2. DB2 concepts

PROCEDURAL DATABASE PROGRAMMING ( PL/SQL AND T-SQL)

Database Foundations. 6-3 Data Definition Language (DDL) Copyright 2015, Oracle and/or its affiliates. All rights reserved.

Database Systems: Design, Implementation, and Management Tenth Edition. Chapter 8 Advanced SQL

RDBMS - PL SQL - Topic 5 - MSBTE QUESTIONS AND ANSWERS

Question: Which statement would you use to invoke a stored procedure in isql*plus?

RESTRICTING AND SORTING DATA

ADVANTAGES. Via PL/SQL, all sorts of calculations can be done quickly and efficiently without use of Oracle engine.

Lab # 4. Data Definition Language (DDL)

INDEX. 1 Introduction. 2 Types of Cursors 2.1 Explicit cursor 2.2 Attributes 2.3 Implicit cursor 2.4 Attributes. 3 Parameterized cursor

What are temporary tables? When are they useful?

Misc. Triggers Views Roles Sequences - Synonyms. Eng. Mohammed Alokshiya. Islamic University of Gaza. Faculty of Engineering

Triggers- View-Sequence

Introduction to Computer Science and Business

Introduction p. 1 The Logical and Physical View of Tables p. 1 Database Types p. 4 NULLs p. 6 DDL and DML Statements p. 7 Column and Table Constraint

CS6312 DATABASE MANAGEMENT SYSTEMS LABORATORY L T P C

CHAPTER4 CONSTRAINTS

To understand the concept of candidate and primary keys and their application in table creation.

D B M G. SQL language: basics. Managing tables. Creating a table Modifying table structure Deleting a table The data dictionary Data integrity

SQL:1999 and Recent Developments in SQL Standardisation

Slides by: Ms. Shree Jaswal

An Introduction to Structured Query Language

Oracle Database 10g: Introduction to SQL

DB2 SQL Class Outline

Where are we? Week -4: Data definition (Creation of the schema) Week -3: Data definition (Triggers) Week -1: Transactions and concurrency in ORACLE.

Vendor: Oracle. Exam Code: 1Z Exam Name: Oracle Database 11g: Program with PL/ SQL. Version: Demo

Sql Server Syllabus. Overview

SQL. Char (30) can store ram, ramji007 or 80- b

RNDr. Michal Kopecký, Ph.D. Department of Software Engineering, Faculty of Mathematics and Physics, Charles University in Prague

An Introduction to Structured Query Language

SQL is an English like language consisting of commands to store, retrieve, maintain & regulate access to your database.

1 Writing Basic SQL SELECT Statements 2 Restricting and Sorting Data

COSC 304 Introduction to Database Systems SQL DDL. Dr. Ramon Lawrence University of British Columbia Okanagan

CSE 530A. Inheritance and Partitioning. Washington University Fall 2013

Databases - 3. Null, Cartesian Product and Join. Null Null is a value that we use when. Something will never have a value

Databases - 3. Null, Cartesian Product and Join. Null Null is a value that we use when. Something will never have a value

Chapter 7. Advanced SQL. Database Systems: Design, Implementation, and Management, Sixth Edition, Rob and Coronel


Relational Database Management Systems Mar/Apr I. Section-A: 5 X 4 =20 Marks

Table of Contents. Oracle SQL PL/SQL Training Courses

ACCESS isn t only a great development tool it s

Data Organization and Processing I

Indexes (continued) Customer table with record numbers. Source: Concepts of Database Management

An Introduction to Structured Query Language

Working with DB2 Data Using SQL and XQuery Answers

Topics Fundamentals of PL/SQL, Integration with PROIV SuperLayer and use within Glovia

ORACLE: PL/SQL Programming

Chapter 14. Active Databases. Database Systems(Part 2) p. 200/286

Real-World Performance Training SQL Introduction

Constraints. Primary Key Foreign Key General table constraints Domain constraints Assertions Triggers. John Edgar 2

2. Programming written ( main theme is to test our data structure knowledge, proficiency

ORACLE Job Placement Paper. Paper Type : General - other

Transcription:

course: Database Applications (NDBI026) WS2015/16 RNDr. Michal Kopecký, Ph.D. Department of Software Engineering, Faculty of Mathematics and Physics, Charles University in Prague

Triggers Triggers Overview Creating triggers Using Triggers Set operators UNION [ALL], INTERSECT, EXCEPT WITH SELECT Statements CASE Expressions M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 2

Active database = a database that can respond to conditions both inside and outside the database. Possible uses Security monitoring, Alerting, Statistics gathering, Authorization ECA rules (Event Condition Action) Triggers M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 4

Trigger = Special type of stored procedure Activated automatically according to events if predefined condition is met on the server Data modification (described later) Events attached to database or session Logging-in/-out Table creation etc. M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 5

Rules Event Condition Action Event data modification (INSERT/UPDATE/DELETE) Condition changed data correspond to condition, defined by Boolean SQL expression Action defined trigger body containing additional INSERT/UPDATE/DELETE statements is executed M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 6

More detailed rules than integrity constraints Implemented directly in the database Applied equally for all client application and sessions Active database allows better enforce broader variety of integrity requirements than passive database, where additional tests have to be implemented on clients and/or application servers. M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 7

Triggers (rules) can be named by any identifies It is advisable to construct names systematically in such a way, that the name suggests the name of the associated table, the operation they react on and other useful information M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 8

Two types of triggers (active database rules) Row-level triggers They are activated by the change of individual rows of the associated table If more rows are changed, they can test data of individual rows Statement level They are activated only once by the action itself independently on the data content and changes M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 9

Detailed checking of changed data, not reachable by integrity constraints Typically by BEFORE row-level triggers Checking that the user is allowed to execute given statement Typically by BEFORE statement-level triggers Journaling of old table values Typically by AFTER row-level triggers Journaling of the operation history on the table Typically by AFTER statement-level triggers Data replication All modifications are repeated on other (remote) table Typically by AFTER row-level triggers M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 10

Addition fixing of application logic Part of the application logic is hidden Changes in data has unpredictable side-effects Worse application debugging It is better to use procedures, that the applications can (or have to due to inability to change data directly) explicitly execute M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 11

CREATE [OR REPLACE] TRIGGER trigger_name {BEFORE AFTER} {INSERT DELETE UPDATE [OF column_name]} ON table_name [FOR EACH ROW] [REFERENCING OLD AS old_name NEW AS new_name] [WHEN (condition)] PL/SQL block; Defines what to be done before, resp. after the change of data in the database Row-level triggers are executed repeatedly for each changed row and see the old and new version of the row Statement-level triggers for the statement and don t see changed data M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 12

Table actualization is done as follows: 1. BEFORE statement-level triggers 2. BEFORE row-level triggers 3. Row actualization / integrity constraint checking 4. AFTER row-level triggers 5. AFTER statement-level triggers If any of those steps fails, ie. The unhandled exception is thrown, actualization is invalid and is not done There can be defined more triggers of the same type on the given table, but their exact order is not defined M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 13

{CREATE ALTER} TRIGGER trigger_name ON { table view } [ WITH ENCRYPTION ] { FOR AFTER INSTEAD OF } { [ INSERT ] [, ] [ UPDATE ] [, ] [ DELETE ] } [ WITH APPEND ] AS [{IF UPDATE ( column ) [{ AND OR } UPDATE ( column ) ] IF ( COLUMNS_UPDATED ( bitwise_operator updated_bitmask )}] sql_statement [...] FOR is older equivalent for AFTER, no statement-level triggers The trigger can see and access all changed data using pseudo-tables DELETED AND INSERTED M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 14

Table actualization is done as follows : 1. Integrity constraint checking 2. Vytvoření tabulek deleted a inserted 2. Aktualizace (INSTEAD OF trigger) / integrity constraint checking 3. AFTER triggers If the transaction is found to be incorrect the transaction can be rolled back using ROLLBACK TRANSACTION The order of execution for triggers of the same type can be defined using sp_settriggerorder procedure M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 15

More types of actualization can be defined for one trigger In Oracle using OR CREATE TRIGGER BEFORE INSERT OR UPDATE ON In the trigger body are defined Boolean predicates INSERTING, UPDATING, DELETING In MS SQL using commas CREATE TRIGGER ON FOR INSERT, UPDATE M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 16

INSTEAD OF triggers overrides the implementation of the data actualization itself They are executed instead of INSERT, UPDATE, DELETE operation In Oracle they are allowed only for views, not for tables MS SQL doesn t distinguish between tables and views M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 17

Scenario I: Checking if the user is allowed to do given operation Executed usually if possible once before the statement is executed It can be tested if the user is logged from internal secure terminal, in working hours and many other things not available using integrity constraints M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 18

Example: CREATE OR REPLACE TRIGGER bef_ins_emp BEFORE INSERT ON Emp BEGIN IF USERENV( TERMINAL ) NOT LIKE PERS% THEN RAISE_APPLICATION_ERROR( -20001, Employees can be hired only from Human resource department ); END IF; END; / M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 19

Example: CREATE TRIGGER aft_ins_emp ON Emp AFTER INSERT AS BEGIN IF HOST_NAME() NOT LIKE PERS% RAISERROR( 'Employees can be hired only from Human resource department', 15, 1); ROLLBACK TRANSACTION; END; M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 20

Scenario II: Computing of functionally dependent values in the table Useful for fast search according to such functional dependent value Not necessary if the database is able to index expressions Executed usually if possible before the actualization on each row M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 21

There are defined two records inside rowlevel triggers NEW and OLD. They contain row data before and after change. INSERT triggers only NEW record DELETE triggers only OLD record UPDATE triggers both OLD and NEW records Names of records can be changed using REFERENCING clause M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 22

Example: CREATE OR REPLACE TRIGGER bef_ins_upd_emp_row BEFORE INSERT OR UPDATE ON Emp FOR EACH ROW BEGIN -- for searching by total income :NEW.Income := COALESCE(:NEW.Sal,0)+COALESCE(NEW.Comm,0); END; / M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 23

Trigger execution can be restricted by a Boolean condition declared in the WHEN condition CREATE OR REPLACE TRIGGER bef_ins_emp_row BEFORE INSERT ON Emp FOR EACH ROW WHEN (NEW.EmpNo IS NULL) -- there is no colon sign in the WHEN clause before NEW and/or OLD BEGIN SELECT SeqEmpNo.nextval INTO :NEW.EmpNo FROM DUAL; END; / M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 24

There are two pseudo-tables visible in the body of each trigger: inserted and deleted that contain new and old version of all changed records deleted in AFTER DELETE, UPDATE trigger inserted in AFTER INSERT, UPDATE triggers Can be tested by: IF EXISTS(SELECT * FROM DELETED) IF EXISTS(SELECT * FROM INSERTED) T/F DELETE, T/T UPDATE, F/T INSERT M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 25

Example: CREATE TRIGGER aft_ins_upd_emp_row ON Emp AFTER INSERT, UPDATE AS BEGIN -- for searching by total income UPDATE Emp SET Income = COALESCE(Sal,0)+COALESCE(Comm,0) WHERE EpmNo IN (SELECT EmpNo FROM inserted) AND (UPDATED(EmpNo) OR UPDATED(Comm)); END; Have to do it after operation using pseudotables deleted and inserted. M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 26

Scenario III: Checking of inserted/updated data Executed if available before the operation and for each row M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 27

Example: CREATE OR REPLACE TRIGGER bef_upd_emp_row BEFORE UPDATE ON Emp FOR EACH ROW BEGIN IF :NEW.Sal > :OLD.Sal * 1.1 THEN RAISE_APPLICATION_ERROR( -21001, Salary can be raised at most by 10%. ); END IF; END; M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 28

Example: CREATE TRIGGER aft_upd_emp_row ON Emp AFTER UPDATE BEGIN IF 0 < ( SELECT COUNT(*) FROM Emp AS NEW INNER JOIN INSERTED AS OLD ON (OLD.EmpNo=NEW.EmpNo) WHERE NEW.Sal > OLD.Sal * 1.1) RAISERROR('Salary can be raised at most by 10%', 15, 1 ); ROLLBACK TRANSACTION; END; Have to do it after operation using pseudo-tables deleted and inserted. M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 29

Scenario IV: Journaling of data changes. Should be done after successful change in the main table. It is possible to remember author of the change, time of change, original and new values etc. M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 30

Example: CREATE OR REPLACE TRIGGER aft_upd_emp_row AFTER UPDATE ON Emp FOR EACH ROW WHEN (OLD.Sal <> NEW.Sal) BEGIN INSERT INTO Journal(Who, When, EmpNo, OldSal, NewSal) VALUES(USER, SYSDATE, :OLD.EmpNo, :OLD.Sal, :NEW.Sal); END; M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 31

Example: CREATE TRIGGER aft_upd_emp_row ON Emp AFTER UPDATE AS BEGIN INSERT INTO Journal(Who, When, EmpNo, OldSal, NewSal) SELECT CURRENT_USER, CURRENT_DATE, d.empno, d.sal, e.sal FROM deleted AS d JOIN employee AS e ON d.emp.id=e.emp_id WHERE UPDATED(Sal); END; M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 32

CREATE [OR REPLACE] TRIGGER trigger_name INSTEAD OF {INSERT DELETE UPDATE [OF column_name]} ON view_name FOR EACH ROW [REFERENCING OLD AS old_name NEW AS new_name] [WHEN (condition)] PL/SQL block; Defines, how the given change should be implemented on underlying tables in the DB M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 33

Views USER_TRIGGERS USER_SOURCE Views SYS.TRIGGERS SYS.TRIGGER_EVENTS M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 34

There are three set operators defined in ANSI SQL UNION [ALL] SELECT FROM WHERE UNION ALL SELECT FROM WHERE UNION ALL If ALL option is not used, the result contains only unique rows, else results are simply concatenated and can contain duplicities Uniqueness of the UNION (and other operators as well) requires either ordering or hashing of operands, if it is possible, it is desirable to avoid unique UNION and use UNION ALL instead. INTERSECT (Unique) intersect of two results of SELECT statement EXCEPT (Oracle uses its proprietary operator MINUS) (Unique) difference of two results of SELECT statement M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 36

Standard SQL contains support for recursive queries since version ANSI SQL-99 in WITH statement Oracle supports it Since version 9i release 2 MS SQL supports it Since version MS SQL Server 2005 Oracle SQL contains also proprietary extension (with limited capabilities) using special clauses START WITH and CONNECT BY in the SELECT statement M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 38

WITH t1[(c11[, ])] AS (SELECT ) [,t2[(c21[, ]) AS (SELECT )] [, ] SELECT FROM t1, t2, WITH construct defines sources t1, t2, used for data selection. Sources t1, t2, are ad-hoc views existing and visible only In the associated SELECT statement Source definitions can (but need not) reference themselves and previous definitions in WITH statement SELECT can (but need not) use defined sources M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 39

Example: Highest number of employees in one department WITH x AS ( SELECT DeptNo, COUNT(*) AS Cnt FROM Emp GROUP BY DeptNo) SELECT MAX(Cnt) AS MaxCnt FROM x; M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 40

Using WITH WITH x AS ( SELECT DeptNo, COUNT(*) AS Cnt FROM Emp GROUP BY DeptNo) SELECT MAX(x.Cnt) FROM x; Without WITH SELECT MAX(x.Cnt) FROM ( SELECT DeptNo, COUNT(*) AS Cnt FROM Emp GROUP BY DeptNo ) x ; M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 41

Using WITH WITH x AS ( SELECT DeptNo, COUNT(*) AS Cnt FROM Emp GROUP BY DeptNo) SELECT a.deptno, b.deptno, a.cnt FROM x AS a, x AS b WHERE a.cnt=b.cnt AND a.deptno<b.deptno; M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 42

Example: Table Emp with following structure: EmpNo Employee ID EName Name of the employee Mgr ID of Employee s manager (direct supervisor) We would like to list all employees together with the degree of subordination (level in the company hierarchy). Employees without supervisor have degree (level) equal to one. M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 43

Example: WITH e(ename,degree) AS ( -- all employees without supervisor (at level one) SELECT x.ename, 1 AS Degree FROM Emp x WHERE Mgr IS NULL UNION ALL -- recursive part -- all employees at level k+1 (supervised by those at level k) SELECT y.ename, e.degree+1 AS Degree FROM Emp y JOIN e ON y.mgr=e.empno ) SELECT * FROM e; M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 44

SELECT columns FROM table_name [WHERE condition3] START WITH condition1 CONNECT BY [NOCYCLE] condition2 [ORDER BY ] [ORDER SIBLINGS BY ] Rows complying the condition1 in START WITH are supposed to be rows at level 1 M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 45

SELECT columns FROM table_name [WHERE condition3] START WITH condition1 CONNECT BY [NOCYCLE] condition2 [ORDER BY ] [ORDER SIBLINGS BY ] For each row found at level k are recursively searched direct descendants at level k+1, that satisfy condition2 in CONNECT BY clause Condition2 has to deal with parent and descendant rows. Values belonging to parent row are preceded by keyword PRIOR NOCYCLE stops recursion if the node should appear second time in the same path. Without it the reappearance causes error. M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 46

SELECT columns FROM table_name [WHERE condition3] START WITH condition1 CONNECT BY [NOCYCLE] condition2 [ORDER BY ] [ORDER SIBLINGS BY ] At the end all rows not satisfying condition3 in WHERE clause are eliminated (descendants were already processed and so they remain in the result even if their predecessor is removed afterwards) M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 47

SELECT columns FROM table_name [WHERE condition3] START WITH condition1 CONNECT BY [NOCYCLE] condition2 [ORDER BY ] [ORDER SIBLINGS BY ] Without ORDER BY the order of rows corresponds to preorder search. Each row contains a pseudocolumn LEVEL, with the level in hierarchy. Order of descendants of one particular parent is not defined by default. It can be specified by the ORDER SIBLINGS BY clause M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 48

Example: SELECT EName, Level AS Degree FROM Emp e START WITH Mgr IS NULL CONNECT BY PRIOR EmpNo = Mgr; SELECT with padding according to level LPAD(,2*Level) EName AS EName, M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 49

It is often necessary to return value where the expression used depends on some value Decoding of value using fixed codebook 1 ~ Monday 7 ~ Sunday 0 ~ 0% VAT, 1 ~ 15% VAT, 2 ~ 21% VAT etc. It can be ineffective to create function for each such situation or define the table for those codebooks and complicate queries by additional join (here the set of values is fixes and doesn t change or change minimal in time) M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 51

ANSI SQL-92 defines two forms of CASE expression a) simple CASE expression: CASE expression WHEN pattern1 THEN value1 WHEN pattern2 THEN value2 [ELSE default] END Returns value for first match expression=pattern i, resp. default value, resp. NULL M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 52

ANSI SQL-92 defines two forms of CASE expression b) searched CASE expression: CASE WHEN condition1 THEN value1 WHEN condition2 THEN value2 [ELSE default] END Return value for first match condition i = TRUE, resp. default value, resp. NULL M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 53

Example: SELECT Course, CASE Day WHEN 1 THEN Monday WHEN 2 THEN Tuesday WHEN 7 THEN Sunday ELSE UNKNOWN DAY END AS Day_of_Week FROM Schedule; M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 54

Example: Sales split according to VAT type: SELECT CASE WHEN TaxGrp=0 THEN Price ELSE NULL END AS Price_00, CASE WHEN TaxGrp=1 THEN Price*1.15 ELSE NULL END AS Price_15, CASE WHEN TaxGrp=2 THEN Price*1.21 ELSE NULL END AS Price_21 FROM Sales; M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 55

Oracle implements additional proprietary function DECODE DECODE(expression pattern1, value1, pattern2, value2[, ][, default]) Corresponds to simple CASE expression Less universal (available only in SELECTs) Not portable M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 56

Example: SELECT Course, DECODE(Day 1, Monday, 2, Tuesday, 7, Sunday UNKNOWN DAY ) AS Day_of_Week FROM Schedule; Example: SELECT Course, CASE Day WHEN 1 THEN Monday WHEN 2 THEN Tuesday WHEN 7 THEN Sunday ELSE UNKNOWN DAY END AS Day_of_Week FROM Schedule; M. Kopecký Data Security - Triggers, WITH SELECT Statements (NDBI026, Lect. 4) 57