Chapter 22: Object-Based

Databases

Database System Concepts, 6th Ed.

©Silberschatz, Korth and Sudarshan
See www.db-book.com for conditions on re-use
 Chapter 22: Object-Based
Databases
 Complex Data Types and Object Orientation
 Structured Data Types and Inheritance in SQL
 Table Inheritance
 Array and Multiset Types in SQL
 Object Identity and Reference Types in SQL
 Implementing O-R Features
 Persistent Programming Languages
 Comparison of Object-Oriented and Object-Relational
Databases

Database System Concepts - 6th Edition 22.2 ©Silberschatz, Korth and Sudarshan
 Object-Relational Data Models
 Extend the relational data model by including object
orientation and constructs to deal with added data
types.
 Allow attributes of tuples to have complex types,
including non-atomic values such as nested relations.
 Preserve relational foundations, in particular the
declarative access to data, while extending modeling
power.
 Upward compatibility with existing relational languages.

Database System Concepts - 6th Edition 22.3 ©Silberschatz, Korth and Sudarshan
 Complex Data Types
 Complex types are nested data structures composed of primitive data
types. These data structures can also be composed of other complex
types. Some examples of complex types include struct(row),
array/list, map and union.

 Complex types are supported by most programming languages

including Python, C++ and Java. They are also supported by
databases like PostgreSQL which introduced composite (struct) type
in 8.0.

Database System Concepts - 6th Edition 22.4 ©Silberschatz, Korth and Sudarshan
 Complex Data Types
 Motivation:
 Permit non-atomic domains (atomic 
indivisible)
 Example of non-atomic domain: set of
integers,or set of tuples
 Allows more intuitive modeling for
applications with complex data
 Intuitive definition:
 allow relations whenever we allow atomic
(scalar) values — relations within relations
 Retains mathematical foundation of relational
model
 Violates first normal form.

Database System Concepts - 6th Edition 22.5 ©Silberschatz, Korth and Sudarshan
 Example of a Nested Relation
 Example: library information system
 Each book has
 title,
 a list (array) of authors,
 Publisher, with subfields name and branch,
and
 a set of keywords
 Non-1NF relation books

Database System Concepts - 6th Edition 22.6 ©Silberschatz, Korth and Sudarshan
 4NF Decomposition of Nested
Relation
 Suppose for simplicity
that title uniquely
identifies a book
 In real world ISBN
is a unique
identifier
 Decompose books
into 4NF using the
schemas:
 (title, author,
position )
 (title, keyword )
 (title, pub-name,
pub-branch )
 4NF design requires
users to include joins
in their queries.

Database System Concepts - 6th Edition 22.7 ©Silberschatz, Korth and Sudarshan
 Complex Types and SQL
 Extensions introduced in SQL:1999 to support complex
types:
 Collection and large object types
 Nested relations are an example of collection
types
 Structured types
 Nested record structures like composite
attributes
 Inheritance
 Object orientation
 Including object identifiers and references
 Not fully implemented in any database system
currently
 But some features are present in each of the major
commercial database systems
 Read the manual of your database system to see
what it supports

Database System Concepts - 6th Edition 22.8 ©Silberschatz, Korth and Sudarshan
 Structured Types and Inheritance
in SQL
 Structured types can be declared and used in SQL
create type Name as
(firstname varchar(20),
lastname varchar(20))
final
create type Address as
(street varchar(20),
city varchar(20),
zipcode varchar(20))
not final
 Note: final and not final indicate whether subtypes can be
created
 Structured types can be used to create tables with composite
attributes
create table person (
name Name,
address Address,
dateOfBirth date)
 Dot notation used to reference components: name.firstname

Database System Concepts - 6th Edition 22.9 ©Silberschatz, Korth and Sudarshan
 Structured Types (cont.)
 User-defined row types
create type PersonType as (
name Name,
address Address,
dateOfBirth date)
not final
 Can then create a table whose rows are a user-defined
type
create table person of PersonType
 Alternative using unnamed row types.
create table person_r(
name row(firstname varchar(20),
lastname varchar(20)),
address row(street varchar(20),
city varchar(20),
zipcode varchar(20)),
dateOfBirth date)

Database System Concepts - 6th Edition 22.10 ©Silberschatz, Korth and Sudarshan
  The following query illustrates how to access
component attributes of a composite attribute.
select name.lastname, address.city
from person

Database System Concepts - 6th Edition 22.11 ©Silberschatz, Korth and Sudarshan
 Methods
 A structure type can have methods defined on it. We
declare methods as part of the type definition of a
structured type.
create type PersonType as (
name Name,
address Address,
dateOfBirth date)
not final
method ageOnDate (onDate date)
returns interval year;

Database System Concepts - 6th Edition 22.12 ©Silberschatz, Korth and Sudarshan
  Method body is given separately.
create instance method ageOnDate (onDate date)
returns interval year
for PersonType
begin
return onDate - self.dateOfBirth;
end
 We can now find the age of each customer by invoking
the method:
select name.lastname, ageOnDate (current_date)
from customer

Database System Concepts - 6th Edition 22.13 ©Silberschatz, Korth and Sudarshan
 Constructor Functions
 Constructor functions are used to create values of structured
types
 A function with the same name as a structured type is a
constructor function.
 E.g.
create function Name(firstname varchar(20), lastname
varchar(20))
returns Name
begin
set self.firstname = firstname;
set self. lastname = lastname;
end
 To create a value of type Name, we use
new Name(‘John’, ‘Smith’)
 Normally used in insert statements
insert into Person values
(new Name(‘John’, ‘Smith),
new Address(’20 Main St’, ‘New York’, ‘11001’),
date ‘1960-8-22’);

Database System Concepts - 6th Edition 22.14 ©Silberschatz, Korth and Sudarshan
 Type Inheritance
 Suppose that we have the following type definition for people:
create type Person
(name varchar(20),
address varchar(20))
 Using inheritance to define the student and teacher types
create type Student
under Person
(degree varchar(20),
department varchar(20))
create type Teacher
under Person
(salary integer,
department varchar(20))
 Subtypes can redefine methods by using overriding method in place of
method in the method declaration

Database System Concepts - 6th Edition 22.15 ©Silberschatz, Korth and Sudarshan
 Multiple Type Inheritance
 SQL:1999 and SQL:2003 do not support multiple inheritance
 If our type system supports multiple inheritance, we can define a type for
teaching assistant as follows:
create type Teaching Assistant
under Student, Teacher
 To avoid a conflict between the two occurrences of department we can
rename them
create type Teaching Assistant
under
Student with (department as student_dept ),
Teacher with (department as teacher_dept )
 Each value must have a most-specific type

Database System Concepts - 6th Edition 22.16 ©Silberschatz, Korth and Sudarshan
 Table Inheritance
 Tables created from subtypes can further be specified as
subtables
 E.g. create table people of Person;
create table students of Student under people;
create table teachers of Teacher under people;
 Tuples added to a subtable are automatically visible to
queries on the supertable
 E.g. query on people also sees students and teachers.
 Similarly updates/deletes on people also result in
updates/deletes on subtables
 To override this behaviour, use “only people” in query
 Conceptually, multiple inheritance is possible with tables
 e.g. teaching_assistants under students and teachers
 But is not supported in SQL currently
 So we cannot create a person (tuple in people) who
is both a student and a teacher

Database System Concepts - 6th Edition 22.17 ©Silberschatz, Korth and Sudarshan
 Consistency Requirements for
Subtables
 Consistency requirements on subtables and supertables.
 Each tuple of the supertable (e.g. people) can
correspond to at most one tuple in each of the
subtables (e.g. students and teachers)
 Additional constraint in SQL:1999:
All tuples corresponding to each other (that is, with
the same values for inherited attributes) must be
derived from one tuple (inserted into one table).
 That is, each entity must have a most specific type
 We cannot have a tuple in people corresponding to
a tuple each in students and teachers

Database System Concepts - 6th Edition 22.18 ©Silberschatz, Korth and Sudarshan
 Array and Multiset Types in SQL
 Example of array and multiset declaration:
create type Publisher as
(name varchar(20),
branch varchar(20));
create type Book as
(title varchar(20),
author_array varchar(20) array [10],
pub_date date,
publisher Publisher,
keyword-set varchar(20) multiset);
create table books of Book;

Database System Concepts - 6th Edition 22.19 ©Silberschatz, Korth and Sudarshan
 Creation of Collection Values
 Array construction
array [‘Silberschatz’,`Korth’,`Sudarshan’]

 Multisets
multiset [‘computer’, ‘database’, ‘SQL’]

 To create a tuple of the type defined by the books relation:

(‘Compilers’, array[`Smith’,`Jones’],
new Publisher (`McGraw-Hill’,`New York’),
multiset [`parsing’,`analysis’ ])

 To insert the preceding tuple into the relation books

insert into books
values
(‘Compilers’, array[`Smith’,`Jones’],
new Publisher (`McGraw-Hill’,`New York’),
multiset [`parsing’,`analysis’ ]);

Database System Concepts - 6th Edition 22.20 ©Silberschatz, Korth and Sudarshan
 Querying Collection-Valued
Attributes
 To find all books that have the word “database” as a keyword,

select title
from books
where ‘database’ in (unnest(keyword-set ))
 We can access individual elements of an array by using indices
 E.g.: If we know that a particular book has three authors, we
could write:
select author_array[1], author_array[2], author_array[3]
from books
where title = `Database System Concepts’
 To get a relation containing pairs of the form “title,
author_name” for each book and each author of the book
select B.title, A.author
from books as B, unnest (B.author_array) as A (author )
 To retain ordering information we add a with ordinality clause
select B.title, A.author, A.position
from books as B, unnest (B.author_array) with ordinality
as
A (author, position )

Database System Concepts - 6th Edition 22.21 ©Silberschatz, Korth and Sudarshan
 Unnesting
 The transformation of a nested relation into a form with
fewer (or no) relation-valued attributes us called
unnesting.
 E.g.
select title, A as author, publisher.name as pub_name,

publisher.branch as pub_branch, K.keyword

from books as B, unnest(B.author_array ) as A
(author ),
unnest (B.keyword_set ) as K (keyword )
 Result relation flat_books

Database System Concepts - 6th Edition 22.22 ©Silberschatz, Korth and Sudarshan
 Nesting
 Nesting is the opposite of unnesting, creating a collection-
valued attribute
 Nesting can be done in a manner similar to aggregation, but
using the function collect() in place of an aggregation
operation, to create a multiset
 To nest the flat_books relation on the attribute keyword:
select title, author, Publisher (pub_name, pub_branch ) as
publisher,
collect (keyword) as keyword_set
from flat_books
groupby title, author, publisher
 To nest on both authors and keywords:
select title, collect (author ) as author_set,
Publisher (pub_name, pub_branch) as publisher,
collect (keyword ) as keyword_set
from flat_books
group by title, publisher

Database System Concepts - 6th Edition 22.23 ©Silberschatz, Korth and Sudarshan
 Nesting (Cont.)
 Another approach to creating nested relations is to use
subqueries in the select clause, starting from the 4NF
relation books4
select title,
array (select author
from authors as A
where A.title = B.title
order by A.position) as author_array,
Publisher (pub-name, pub-branch) as publisher,
multiset (select keyword
from keywords as K
where K.title = B.title) as keyword_set
from books4 as B

Database System Concepts - 6th Edition 22.24 ©Silberschatz, Korth and Sudarshan
 Object-Identity and Reference
Types
 Define a type Department with a field name and a field head
which is a reference to the type Person, with table people as
scope:
create type Department (
name varchar (20),
head ref (Person) scope people)
 We can then create a table departments as follows
create table departments of Department
 We can omit the declaration scope people from the type
declaration and instead make an addition to the create table
statement:
create table departments of Department
(head with options scope people)
 Referenced table must have an attribute that stores the
identifier, called the self-referential attribute
create table people of Person
ref is person_id system generated;

Database System Concepts - 6th Edition 22.25 ©Silberschatz, Korth and Sudarshan
 Initializing Reference-Typed Values
 To create a tuple with a reference value, we can first
create the tuple with a null reference and then set the
reference separately:
insert into departments
values (`CS’, null)
update departments
set head = (select p.person_id
from people as p
where name = `John’)
where name = `CS’

Database System Concepts - 6th Edition 22.26 ©Silberschatz, Korth and Sudarshan
 User Generated Identifiers
 The type of the object-identifier must be specified as part of
the type definition of the referenced table, and
 The table definition must specify that the reference is user
generated
create type Person
(name varchar(20)
address varchar(20))
ref using varchar(20)
create table people of Person
ref is person_id user generated
 When creating a tuple, we must provide a unique value for the
identifier:
insert into people (person_id, name, address )
values
(‘01284567’, ‘John’, `23 Coyote Run’)
 We can then use the identifier value when inserting a tuple
into departments
 Avoids need for a separate query to retrieve the identifier:

insert into departments

values(`CS’, `02184567’)
Database System Concepts - 6th Edition 22.27 ©Silberschatz, Korth and Sudarshan
 User Generated Identifiers (Cont.)

 Can use an existing primary key value as the identifier:

create type Person
(name varchar (20) primary key,
address varchar(20))
ref from (name)
create table people of Person
ref is person_id derived
 When inserting a tuple for departments, we can then use
insert into departments
values(`CS’,`John’)

Database System Concepts - 6th Edition 22.28 ©Silberschatz, Korth and Sudarshan
 Path Expressions
 Find the names and addresses of the heads of all
departments:
select head –>name, head –>address
from departments
 An expression such as “head–>name” is called a path
expression
 Path expressions help avoid explicit joins
 If department head were not a reference, a join of
departments with people would be required to get at
the address
 Makes expressing the query much easier for the user

Database System Concepts - 6th Edition 22.29 ©Silberschatz, Korth and Sudarshan
 Implementing O-R Features
 Similar to how E-R features are mapped onto relation
schemas
 Multivalued attributes correspond to multiset-valued
 Composite attributes correspond to structured type
 ISA hierarchies correspond to table inheritance.
 Subtable implementation
 Each table stores primary key and those attributes
defined in that table
or,
 Each table stores both locally defined and inherited
attributes

Database System Concepts - 6th Edition 22.30 ©Silberschatz, Korth and Sudarshan
 Persistent Programming
Languages
 Persistent data – data that continue to exist even after
the program that created it has terminated.
 Tuples in a relation - DBMS
 Files – traditional prrogramming languages
 The traditional way to access persistent data is by
embedding SQL within the programming language.
 A Persistent programming language is a programming
language extended with constructs to handle persistent
data.

Database System Concepts - 6th Edition 22.31 ©Silberschatz, Korth and Sudarshan
 Languages with embedded Persistent programming
SQL languages
 The programmer is  The Query language and
responsible for any type the host language share
conversions between the same type system.
host language and SQL.  Objects can be created
 Type conversion from and stored in the
relational format to database without any
object-oriented format explicit type or format
takes a substantial changes.
amount of code.  Format changes are
 Format changes are not transparent.
transparent  Programmer can
 Programmer has to write manipulate persistent
code explicitly to fetch data without writing
persistent data into code explicitly to fetch
memory and store it it into memory and store
back to disk it back to disk
Database System Concepts - 6th Edition 22.32 ©Silberschatz, Korth and Sudarshan
 Languages with embedded Persistent programming
SQL languages
 The programmer is  The Query language and
responsible for any type the host language share
conversions between the same type system.
host language and SQL.  Objects can be created
 Type conversion from and stored in the
relational format to database without any
object-oriented format explicit type or format
takes a substantial changes.
amount of code.  Format changes are
 Format changes are not transparent.
transparent  Programmer can
 Programmer has to write manipulate persistent
code explicitly to fetch data without writing
persistent data into code explicitly to fetch
memory and store it it into memory and store
back to disk it back to disk
Database System Concepts - 6th Edition 22.33 ©Silberschatz, Korth and Sudarshan
  C++ and JAVA can be extended to make them persistent
programming languages.
 Persistent objects:
 Persistence by class - explicit declaration of persistence
 Declare that a class is persistent
 All objects of that class are then persistent
 Not flexible, since it is useful to have both transient
and persistent objects in a class
 Persistence by creation - special syntax to create
persistent objects
 Thus an object is either persistent or transient.
 Persistence by marking - make objects persistent after
creation
 All objects are created as transient, but, if an object is
to persist, it is marked as persistent before the
program terminates.

Database System Concepts - 6th Edition 22.34 ©Silberschatz, Korth and Sudarshan
  Persistence by reachability - object is persistent if it
is declared explicitly to be so or is reachable from a
persistent object
 It is easy to make entire data structures
persistent by merely declaring root of such
structures as persistent

Database System Concepts - 6th Edition 22.35 ©Silberschatz, Korth and Sudarshan
 Object Identity and Pointers
 When a persistent object is created , it is assigned a
persistent object identifier.
 The association of an object with a physical location in
storage may change over time.
 Degrees of permanence of object identity
 Intraprocedure: only during execution of a single
procedure. Eg: local variables
 Intraprogram: only during execution of a single
program or query. Eg: global variables
 Interprogram: across program executions, but not if
data-storage format on disk changes
 Persistent: interprogram, plus persistent across data
reorganizations.

Database System Concepts - 6th Edition 22.36 ©Silberschatz, Korth and Sudarshan
  Storage and access of persistent objects:
Name to objects
Expose persistent pointers to objects
Store collection of objects
Class extent

 Persistent versions of C++ and Java have been

implemented
 C++
 ODMG C++
 ObjectStore
 Java
 Java Database Objects (JDO)

Database System Concepts - 6th Edition 22.37 ©Silberschatz, Korth and Sudarshan
 Persistent C++ Systems
 Extensions of C++ language to support persistent storage of
objects
 Several proposals, ODMG standard proposed, but not much
action of late
 persistent pointers: e.g. d_Ref<T> to represent persistent
pointers to a class T.
 creation of persistent objects: e.g. new (db) T() to create
persistent object.
 Class extents: access to all persistent objects of a
particular class
 Relationships: Represented by pointers stored in related
objects
 Iterator interface
 Transactions
 Updates: mark_modified() function to tell system that a
persistent object that was fetched into memory has been
updated
 Query language

Database System Concepts - 6th Edition 22.38 ©Silberschatz, Korth and Sudarshan
 Persistent Java Systems
 Standard for adding persistence to Java : Java Database
Objects (JDO)
 Persistence by reachability
 Byte code enhancement
 Classes separately declared as persistent
 Byte code modifier program modifies class byte code to
support persistence
– E.g. Fetch object on demand
– Mark modified objects to be written back to database
 Database mapping
 Allows objects to be stored in a relational database
 Class extents
 Single reference type
 no difference between in-memory pointer and persistent
pointer
 Implementation technique based on hollow objects
(a.k.a. pointer swizzling)

Database System Concepts - 6th Edition 22.39 ©Silberschatz, Korth and Sudarshan
 Object-Relational Mapping
 Object-Relational Mapping (ORM) systems built on top of
traditional relational databases
 Implementor provides a mapping from objects to
relations
 Objects are purely transient, no permanent object
identity
 Objects can be retried from database
 System uses mapping to fetch relevant data from
relations and construct objects
 Updated objects are stored back in database by
generating corresponding update/insert/delete
statements
 Provides query language operating direcly on object
model
 queries translated to SQL
 Limitations: overheads, especially for bulk updates

Database System Concepts - 6th Edition 22.40 ©Silberschatz, Korth and Sudarshan
 Comparison of O-O and O-R
Databases

 Relational systems
 simple data types, powerful query languages, high
protection.
 Persistent-programming-language-based OODBs
 complex data types, integration with programming
language, high performance.
 Object-relational systems
 complex data types, powerful query languages, high
protection.
 Object-relational mapping systems
 complex data types integrated with programming language,
but built as a layer on top of a relational database system
 Note: Many real systems blur these boundaries
 E.g. persistent programming language built as a wrapper
on a relational database offers first two benefits, but may
have poor performance.

Database System Concepts - 6th Edition 22.41 ©Silberschatz, Korth and Sudarshan
End of Chapter 22

Database System Concepts, 6th Ed.

©Silberschatz, Korth and Sudarshan
See www.db-book.com for conditions on re-use
 Figure 22.05

Database System Concepts - 6th Edition 22.43 ©Silberschatz, Korth and Sudarshan
 Figure 22.07

Database System Concepts - 6th Edition 22.44 ©Silberschatz, Korth and Sudarshan