Beruflich Dokumente
Kultur Dokumente
surendersingh@rediffmail.com
Surender Singh
Sr. Programmer
surendersingh@titsbhiwani.ac.in
surendersingh@rediffmail.com
Relational Database Management System
DATA
DATABASE
DBMS/RDBMS
Information
surendersingh@rediffmail.com
File Processing System
surendersingh@rediffmail.com
File Processing System
Application
Programs
File System
(Programs Database
Written in C (Data Structure (Information in
Pascal etc.) File Handling) Files Format)
surendersingh@rediffmail.com
File System
Database
surendersingh@rediffmail.com
Disadvantages of FPS
surendersingh@rediffmail.com
Data Redundancy and Inconsistency
surendersingh@rediffmail.com
Difficulty in accessing data
Manager
Requirement
Application
Programs
File System
(Programs Database
Written in C (Data Structure (Information Storage
Pascal etc.) File Handling) in Files Format)
surendersingh@rediffmail.com
Data Isolation and Integrity Problems
Program in C Program in COBOL
#include <stdio.h>
01 Reserve-rec.
Main()
03 saving
{
05 accno PIC A(2)
-----
--------
}
New Document
surendersingh@rediffmail.com
Atomicity Problems
Bank
Data Transmission
surendersingh@rediffmail.com
USER USER
Concurrent-access anomalies
surendersingh@rediffmail.com
Security Problems
Employee
Information
surendersingh@rediffmail.com
Database the Piece of mind
surendersingh@rediffmail.com
Requirements of a DBMS
surendersingh@rediffmail.com
A DBMS has two major components, namely
D
B File
Secondary
M Manager
Storage
S
surendersingh@rediffmail.com
View of DATA
surendersingh@rediffmail.com
Data Independence
The ability to modify a schema definition in one level without affecting a
schema definition in the next higher level is called data independence.
surendersingh@rediffmail.com
Data Models
A Data Model is a mechanism for describing the data, their interrelationships
and the constraints.
Record-based models.
Relational Model
Network Model
Hierarchical Model
surendersingh@rediffmail.com
The E-R Model
Entities : An entity is a distinct clearly identifiable object of the database e.g Book
Attribute : Each Entity is characterized by a set of attributes e.g. Acc.No.
Entity Set : Set of all entities having attributes of the same type.
Relationships : A relationship is a mapping between entity sets.
surendersingh@rediffmail.com
Author YearofPub Card_No DOI Address
The Relational Model
Relational Model uses a collection of tables to represent both data and
relationship among those data. Each table has multiple Attributes and
similar kind of tuples.
Attribute
Book Table/Relation
Tuple
surendersingh@rediffmail.com
Network Model
Data in the network model are represented by collection of records and
relationships among data are represented by links, which can be viewed
as Pointers.
User
Card_No Name Address Link
Pointer Next
Book
Acc_No Author ----- Link
surendersingh@rediffmail.com
Hierarchical Model
This is special kind of a network model where the relationship is
essentially a tree-like structure.
Hospital
Wards Units
surendersingh@rediffmail.com
Physical Data Models
Physical data models are used to describe data at the lowest level.
In contrast to logical data models, there are few physical data models
In use. Two of the widely known ones are the Unifing model and
frame-Memory model.
surendersingh@rediffmail.com
Database Languages
Database Languages
Data-Definition Data-Manipulation Data-Control
surendersingh@rediffmail.com
Database Management System Structure
Nave Users Application Sophisticated Database
Users
(tellers, agents, etc.) Programmers Users Administrators
Embedded
DML DDL
Application DML
Compiler Interpreter
Programs Precompiler
Object Code Query
Query
Processor
Evaluation
Engine
Database
Management
System
Transaction Buffer
Manager
Storage
Manager
Manager
File
Manager
surendersingh@rediffmail.com
Data Files Data Dictionary
surendersingh@rediffmail.com
Oracle Storage System Structure
surendersingh@rediffmail.com
Database Administrator
Roles of DBA
Schema Definition
Storage structure and access-method definition
Schema and Physical-organization modification
Granting of authorization for data access
Integrity-constraint specification
surendersingh@rediffmail.com
Terms
Simple and Composite Attributes
Single-valued and Multivalued Attributes
Null Attributes
Derived Attributes
Existence Dependencies
Weak Entity Set and Strong Entity Set
surendersingh@rediffmail.com
Weak Entity Set
surendersingh@rediffmail.com
Attributes
surendersingh@rediffmail.com
Keys
Keys
Composite Key
surendersingh@rediffmail.com
Candidate Keys
surendersingh@rediffmail.com
Primary Secondary Key
Key
Roll_No Name Branch City
01 Deepak Computers Bhiwani
02 Mukesh Electronics Rohtak
03 Teena Computers Bhiwani
04 Deepak Electronics Rohtak
05 Monika Computers Delhi
surendersingh@rediffmail.com
Composite Primary Key
surendersingh@rediffmail.com
P#
Part P_Name Colour Quantity
P1 Nut Red 200
P2 Bolt Green 250
P3 Screw Blue 300
surendersingh@rediffmail.com
S#
Supplier S_Name City Quantity
S1 John Delhi 200
S2 Smith Kolkata 250
S3 James Delhi 300
S4 David Chennai 400
S5 John Chennai 300
surendersingh@rediffmail.com
SP#
P# S# Quantity
P1 S1 200
P2 S1 300
P3 S1 400
P1 S2 250
P2 S3 250
P3 S4 200
P2 S4 300
P3 S5 400
surendersingh@rediffmail.com
Mapping Cardinalities
For a binary relationship set R between entity sets A and B, the mapping
Cardinality must be one of the following
A B A B
surendersingh@rediffmail.com
A B A B
surendersingh@rediffmail.com
Many to One Many to Many
More on E-R Diagrams
Company
Vehicle
Manager
Staff Reports to
Subordinate
Circular Relationship
surendersingh@rediffmail.com
Ternary E-R Diagram
Courses
Constraints
surendersingh@rediffmail.com
E-R Diagram Components
Entity Sets
Attributes
Relationship Sets
Connectors/Constraints
Multivalued Attributes
Derived Attributes
surendersingh@rediffmail.com
Existence Dependencies
surendersingh@rediffmail.com
Generalization and Specialization
surendersingh@rediffmail.com
Generalization and Specialization
The abstraction mechanisms
IS_A IS_A
Full_time Part_time
Type
Employee Salary Employee
surendersingh@rediffmail.com
Degree Interest Stipend Hour_Rate
Aggregation
The Process of compiling information on an object
Teaches
Book
Teacher-Teaches
Uses
surendersingh@rediffmail.com
Book
Represent ER model using tables
surendersingh@rediffmail.com
Query Languages
Procedural, where the user instructs the system to perform a sequence of operations
on the database. This will compute the desired information.
Nonprocedural, where the user species the information desired without giving a
procedure for ob-taining the information.
A complete query language also contains facilities to insert and delete tuples as well
as to modify parts of existing tuples.
surendersingh@rediffmail.com
The Relational Algebra
The relational algebra is a procedural query language.
surendersingh@rediffmail.com
Fundamental Operations
select (unary)
project (unary)
rename (unary)
cartesian product (binary)
union (binary)
set-difference (binary)
surendersingh@rediffmail.com
Formal Definition of Relational Algebra
surendersingh@rediffmail.com
The Select Operation
surendersingh@rediffmail.com
The Project Operation
surendersingh@rediffmail.com
The Cartesian Product Operation
surendersingh@rediffmail.com
Output of Cartesian Product
A B A B
1 1 X
X
2 1 Y
Y 2 X
3
2 Y
3 X
3 Y
surendersingh@rediffmail.com
The Rename Operation
surendersingh@rediffmail.com
The Union Operation
surendersingh@rediffmail.com
The Set Difference Operation
surendersingh@rediffmail.com
Additional Operations
surendersingh@rediffmail.com
The Natural Join Operation
surendersingh@rediffmail.com
The Division Operation
surendersingh@rediffmail.com
Example of Division Operation
Relation R Relation S S
R
A B B A
P A A P
Q A
P B B Q
Q T
M A
Q B
surendersingh@rediffmail.com
The Assignment Operation
surendersingh@rediffmail.com
Relational Calculus
surendersingh@rediffmail.com
Tuple Relational Calculus
surendersingh@rediffmail.com
Example Queries
surendersingh@rediffmail.com
Some More Examples
surendersingh@rediffmail.com
Domain Relational Calculus
surendersingh@rediffmail.com
SQL
surendersingh@rediffmail.com
Integrity Constraints
Insertion
Deletion
Updation
Recovery from Failure
Concurrent Operations
Types of Constraints
Domain Constraints
Referential Integrity Constraint
Functional Dependencies
surendersingh@rediffmail.com
Domain Constraints
Includes
Type
Width
Null or Not Null
Checks/Conditions
Specify at the time of designing
Checked at the time of insertion, deletion or modification
e.g
Bname char(20)
Amount number(7,2)
DOL date check (date>=29/09/2004
City char(10) not null
TotalAmt = amount + interest
surendersingh@rediffmail.com
Referential Integrity
Foreign Key
Referential integrity states that all values of the foreign key of one
Relation must be present in another relation where the same attribute
Is declared as the primary key
surendersingh@rediffmail.com
Assertions and Triggers
An assertion is a general predicate, expressed in relational algebra
Or calculus or any language like SQL which must always hold in a
Database
surendersingh@rediffmail.com
Functional Dependencies
Functional Dependencies provide a formal mechanism to express
Constraints between attributes.
surendersingh@rediffmail.com
Formal Notation of FD
In general if there are two attributes A and B and the FD
A B
Holds then, it means that there can be no two tuple which have
The same value of attributes A and different values in attribute B.
1. , R, i.e. , subset of R
2. for all tuples t1 and t2 in r,
if t1 [ ] = t2 [ ] then
t1 [ ] = t2 [ ]
surendersingh@rediffmail.com
Closure of a Set of Functional
Dependencies
surendersingh@rediffmail.com
Armstrongs Axioms
surendersingh@rediffmail.com
Closure of a Set of F+
surendersingh@rediffmail.com
Closure of Attribute Sets
surendersingh@rediffmail.com
Canonical Cover
surendersingh@rediffmail.com
Example of Cannonical Cover
Consider a relation r ( X, Y, Z ) with the FDs F.
1. X YZ
2. Y Z
3. X Y
4. XY Z
Here 4 is redundant because (1) states that X Y and X Z holds.
Thus (4) can be derived from (1). Also (3) is redundant because (1) contains (3).
Deleting these two we get
1. X YZ
2. Y Z
Which is a cover of F. Here again since X Y and Y Z holds, by
Transitivity X Z holds. So it is redundant. Deleting this we get the FDs as
X Y
Y Z
Which is a cannonical cover of F.
surendersingh@rediffmail.com
Relational Database Design
surendersingh@rediffmail.com
Database Decomposition 1
Representation of Information
surendersingh@rediffmail.com
Database Decomposition 2
surendersingh@rediffmail.com
Database Decomposition 3
surendersingh@rediffmail.com
Database Decomposition 4
surendersingh@rediffmail.com
Lossless-join Decomposition
surendersingh@rediffmail.com
Example of lossy decomposition
S_by
s_name s_addr Item Price
A1 B1 C1 D1
A1 B1 C2 D1
p1 A2 B2 C1 D2 p2
S_addr Item price
S_name Item A2 B2 C3 D3
B1 C1 D1
A1 C1 A3 B1 C2 D2
B1 C2 D1
A1 C2
A2 C1 Natural Join of P1 and p2 B2 C1 D2
S_name S_addr Item Price B2 C3 D3
A2 C3
A1 B1 C1 D1 B1 C2 D2
A3 C2
A1 B2 C1 D2
A1 B1 C2 D1
A1 B1 C2 D2
A2 B1 C1 D1
A2 B2 C1 D2
A2 B2 C3 D3
A3 B1 C2 D1
surendersingh@rediffmail.com
A3 B1 C2 D1
Dependency Preservation
surendersingh@rediffmail.com
Normalization
surendersingh@rediffmail.com
First Normal Form (1NF)
Name C_addr
surendersingh@rediffmail.com
Second Normal Form (2NF)
A relation is said to be in 2NF if it is in 1NF and
All non-prime attributes are fully functionally dependent on candidate key
Here [name, acc_no ] is the candidate key and addr and amt are the non prime attributes.
Among the non-prime attributes amt depends on [name, acc_no ] whereas addr depends
on name only.
Note that due to FD name addr every tuple with the same name will contain the same
Address causing redundancy.
This redundancy arises because a non-prime attribute like address is dependent on an attribute
surendersingh@rediffmail.com
Which is not a candidate key.
Solution
We can remove this redundancy by splitting the original relation into following two relations
In the second relation the only non-prime attribute amt depend on both name and
Acc_no. that this decomposition is also lossless join and dependency preserving
Course_no title
Course_no, time loc
surendersingh@rediffmail.com
Third Normal Form (3NF)
A relation is said to be in 3NF and non-prime attributes are not dependent
On each other.
Here all prime attributes are fully functional dependent on candidate keys, the
Non-prime attribute gift-item is also fully functional dependent on the non-prime
Attribute price. This create redundancy because every price value there is a fixed
Gift item.
We shall have to impose the additional restriction that no non-prime attribute can
Be functionally dependent on another non-prime attributes.
surendersingh@rediffmail.com
Solution
We decompose the relation
s_by (s_name, item, price, gift_item )
Into
s_by_1 (s_name, item, price )
s_by_2 (price, gift_item)
(trivial dependency)
R ( is a super key )
surendersingh@rediffmail.com
Boyce-Codd Normal Form (BCNF)
surendersingh@rediffmail.com
More on BCNF
surendersingh@rediffmail.com
Comparison of BCNF and 3NF
surendersingh@rediffmail.com
Comparison of BCNF and 3NF - 2
surendersingh@rediffmail.com
Normalization using Multivalued
Dependencies
surendersingh@rediffmail.com
Multivalued Dependencies -2
surendersingh@rediffmail.com
Rules
surendersingh@rediffmail.com
More Rules
surendersingh@rediffmail.com
Fourth Normal Form (4NF)
surendersingh@rediffmail.com
Example
surendersingh@rediffmail.com
Normalization using Join Dependencies
Let R be a relation schema and R1, R2,.Rn be a decomposition of R. The join dependency
*(R1, R2,.Rn) is used to restrict the set of legal relations to those for which R1, R2,.Rn is
A lossless-join decomposition of R.
surendersingh@rediffmail.com
Fifth Normal Form (5NF)
Project-Join Normal Form
Project-join normal form (PJNF) is defined in a manner similar to BCNF and 4NF,
Except that join dependencies are used.
surendersingh@rediffmail.com
Storage and File Structure
Hierarchy of Storage
surendersingh@rediffmail.com
Description
surendersingh@rediffmail.com
Description - 2
surendersingh@rediffmail.com
File Organization
surendersingh@rediffmail.com
Fixed Length Record -1
surendersingh@rediffmail.com
Fixed Length Record -2
surendersingh@rediffmail.com
Variable-length Records
surendersingh@rediffmail.com
Fixed-length representation
surendersingh@rediffmail.com
Organization of Records in files
surendersingh@rediffmail.com
Concurrency Control and Recovery
Transactions
Concurrent execution of user programs is essential for good DBMS performance.
Because disk accesses are frequent, and relatively slow, it is important to keep the cpu humming by
working on several user programs concurrently.
A users program may carry out many operations on the data retrieved from the database, but the
DBMS is only concerned about what data is read/written from/to the database.
A transaction is the DBMSs abstract view of a user program: a sequence of reads and writes.
A Tracnsaction is a unit of program execution That accesses and possibly updates various
Data items.
Collection of operations that form a single logical unit of work are called tracsactions.
A database system must ensure proper execution of transaction despite failures.
To ensure integrity of the data, database system must maintain the following properties of the
transactions:
surendersingh@rediffmail.com
States of Transactions
Partially Committed
Active
Aborted
Failed
surendersingh@rediffmail.com
Concurrency in a DBMS
Users submit transactions, and can think of each transaction as executing by itself.
Concurrency is achieved by the DBMS, which interleaves actions (reads/writes of DB objects) of
various transactions.
Each transaction must leave the database in a consistent state if the DB is consistent when the
transaction begins.
DBMS will enforce some ICs, depending on the ICs declared in CREATE TABLE statements.
Beyond this, the DBMS does not really understand the semantics of the data. (e.g., it does not
surendersingh@rediffmail.com
Example
surendersingh@rediffmail.com
Example (Contd.)
A
T1 T2 Dependency graph
B
Dependency graph: One node per Xact; edge from Ti to Tj if Tj reads or writes an object last
written by Ti.
The cycle in the graph reveals the problem. The output of T1 depends on T2, and vice-versa.
surendersingh@rediffmail.com
Scheduling Transactions
Equivalent schedules: For any database state, the effect (on the set of objects in the database) of
executing the first schedule is identical to the effect of executing the second schedule.
Serializable schedule: A schedule that is equivalent to some serial execution of the transactions.
If the dependency graph of a schedule is acyclic, the schedule is called conflict serializable. Such a
schedule is equivalent to a serial schedule.
This is the condition that is typically enforced in a DBMS (although it is not necessary for
serializability).
surendersingh@rediffmail.com
Detection of Serializability
One of the techniques of concurrency control is to detect whether a schedule is valid or not
Prior to execution.
The task of understanding a schedule is simplified by considering only the sequence of read
and write operation in a transaction
T1 T2
Read(X)
Read(X)
Write(X)
Write(X)
Read(Y)
Write(Y)
Read(Y)
Write(Y)
surendersingh@rediffmail.com
Serializable Concurrency
T1 T2
Read(X)
Write(X)
Read(X)
Write(X)
Read(Y)
Write(Y)
Read(Y)
Write(Y)
Generalize the idea of conflict. Consider the four possibilities which can arise between two
Consecutive instructions T1 and T2 in a schedule ( T1 and T2 belong to two different transactions)
T1 and T2 are said to be conflict if they cannot be swapped without fear of loss of consistency.
surendersingh@rediffmail.com
In above 3 cases all pairs except case 2 are said to be in conflict.
Deadlock Condition
T1 T2
surendersingh@rediffmail.com
Lock-Based Techniques
In this technique the system does not participate in detection of inconsistency nor does it take any
Corrective action.
The DBMS however, provides the user with a set of operations which when used properly can
ensure that concurrent execution will not violate consistency.
In this techniques functions are provided to lock and unlock data items by transactions,
In the simplest case a data item X can be locked by a transaction T1 in two modes :
Shared Mode : if T1 locks X in shared mode then before T1 unlocks X, no other transaction T2
can write into X. But a transaction T2 can read the value of X even if T1 has locked
locked X in shared mode.
Exclusive Mode : If T1 locks X in exclusive mode then before T1 unlocks X, no other transaction
T2 can read or write into X.
surendersingh@rediffmail.com
Example
T1 T2
Lock-X(P)
Read (P,p)
P=p-1
Write(P,p)
Unlock(P)
Lock-S(Q)
Read(Q,q)
unlock(Q)
Lock-S(P)
Read(P,p)
unlock(P)
display(p)
display(p)
Lock-X(Q)
Read(Q,q)
q=q+1
Write(Q,q)
Unlock(Q)
surendersingh@rediffmail.com
Two-Phase locking
Phase I Acquiring Phase : During this phase a transaction may lock a data item but not
unlock any data item.
Phase II Releasing Phase : During this phase a transaction may unlock data items locked
earlier but no new locks may be acquired.
In two phase locking phase I must always precede phase II. This will ensure that all schedule
are automatically conflict serialzable.
surendersingh@rediffmail.com
Enforcing (Conflict) Serializability
If an Xact holds an X lock on an object, no other Xact can get a lock (S or X) on that object.
2PL allows only conflict-serializable schedules.
Potential problem of deadlocks: we could have a cycle of Xacts, T1, T2, ... , Tn, with each Ti waiting for its
predecessor to release some lock that it needs.
Dealt with by killing one of them and releasing its locks.
surendersingh@rediffmail.com
Atomicity of Transactions
A transaction might commit after completing all its actions, or it could abort (or be aborted by the DBMS)
after executing some actions.
A very important property guaranteed by the DBMS for all transactions is that they are atomic. That is, a
user can think of a Xact as always executing all its actions in one step, or not executing any actions at all.
DBMS logs all actions so that it can undo the actions of aborted transactions.
This ensures that if each Xact preserves consistency, every serializable schedule preserves consistency.
surendersingh@rediffmail.com
Aborting a Transaction
If a transaction Ti is aborted, all its actions have to be undone. Not only that, if Tj reads an object last
written by Ti, Tj must be aborted as well!
Most systems avoid such cascading aborts by releasing a transactions locks only at commit time.
If Ti writes an object, Tj can read this only after Ti commits.
In order to undo the actions of an aborted transaction, the DBMS maintains a log in which every write is
recorded. This mechanism is also used to recover from system crashes: all active Xacts at the time of the
crash are aborted when the system comes back up.
surendersingh@rediffmail.com
The Log
Log records are chained together by Xact id, so its easy to undo a specific Xact.
Log is often duplexed and archived on stable storage.
All log related activities (and in fact, all activities such as lock/unlock, dealing with deadlocks etc.) are
handled transparently by the DBMS.
surendersingh@rediffmail.com
The Log - 2
Log file e.g. X=1000, Y= 2000
T:
Read (X, xi) Transaction Name
xi xi 500 Data item Name
Write (X,xi) Old Value
New Value
Read ( Y, yi)
yi yi + 500 <T starts>
Write (Y, yi) <T, X, 1000, 500>
<T, Y, 2000, 2500>
<T, commits>
surendersingh@rediffmail.com
Checkpoints
At the time of recovery the entire log needs to be searched to know which transaction need to
Be redone and which transactions needs to be undone. The problem with this approach is:
To solve this problem the concept of checkpoint is used here at different points.
Checkpoints are introduced to indicate that the data before this point has already been
Updated to the database. Before writing checkpoints the following sequence of actions
shuld to take place
- Output all log records currently residing in the main store to a stable storage
- Output all modified buffer blocks to secondary storage.
- Output a log record <checkpoint>
surendersingh@rediffmail.com
Recovering From a Crash
There are 3 phases in the Aries recovery algorithm:
Analysis: Scan the log forward (from the most recent checkpoint) to identify all Xacts that were active,
and all dirty pages in the buffer pool at the time of the crash.
Redo: Redoes all updates to dirty pages in the buffer pool, as needed, to ensure that all logged
updates are in fact carried out and written to disk.
Undo: The writes of all Xacts that were active at the crash are undone (by restoring the before value
of the update, which is in the log record for the update), working backwards in the log. (Some care
must be taken to handle the case of a crash occurring during the recovery process!)
Data can be lost due to the failure of the nonvolatile storage like the disk. The scheme which is available
To protect the data from disk failure is to periodically dump the entire contents of the database to any backup
(or even stable) storage like a magnetic tape. When a failure occurs the most recent dump is used to restoring
The datbase to a previous consistent state. Then the log is used to redo all the transactions that have committed
Since the last dump occurred. The following steps are performed for this purpose :
Output all log records currently residing in the main memory onto stable store.
Output all buffer blocks onto the disk.
Copy the contents of the database to stable store.
Output a log record <dump>.
surendersingh@rediffmail.com
Summary
Concurrency control and recovery are among the most important functions provided by a DBMS.
Users need not worry about concurrency.
System automatically inserts lock/unlock requests and schedules actions of different Xacts in such a
way as to ensure that the resulting execution is equivalent to executing the Xacts one after the other in
some order.
Write-ahead logging (WAL) is used to undo the actions of aborted transactions and to restore the
system to a consistent state after a crash.
Consistent state: Only the effects of commited Xacts seen.
surendersingh@rediffmail.com
Query Processing/Optimization
surendersingh@rediffmail.com
Rules
surendersingh@rediffmail.com
Example
surendersingh@rediffmail.com
Example contd.
surendersingh@rediffmail.com
Projection Operation
surendersingh@rediffmail.com
Natural Join Operation
surendersingh@rediffmail.com
Natural Join Operation - 2
surendersingh@rediffmail.com