Transcript Data Models - La Salle University

Data Models
Based in part on Chapter 2 in
Database Systems (Rob and
Coronel)
CSC 240 (Blum)
1
Degrees of Separation
• In the early 1970’s, the Data Base Task
Group (DBTG) identified two levels
important to distinguish in database design.
– The schema is the logical design of the entire
database.
– The sub-schema is the logical design of part of
the database seen by a particular user or
application (a view).
• Codd, Rule 6
CSC 240 (Blum)
2
Schema vs. Instance
• A database schema (its design) should be
distinguished from a database instance, which
also includes the actual data at any given time.
– Analogy: Schema is to instance as class (template) is to
object (instantiation)
• Similarly, the Data Definition Language (DDL) is
used to create/modify the schema, while the Data
Manipulation Language (DML) is mainly used to
modify or retrieve aspects of the instance.
CSC 240 (Blum)
3
ANSI-SPARC Architecture
• In the mid-1970’s, American National
Standards Institute (ANSI) put together the
Standards Planning and Requirements
Committee (SPARC).
• ANSI-SPARC identified three levels
important to distinguish in database design.
– External
– Conceptual
– Internal
CSC 240 (Blum)
4
ANSI web site
CSC 240 (Blum)
5
ANSI-SPARC
External
level
Conceptual
level
Internal
level
View 1
View 2
View 3
Conceptual
Schema
Internal
Schema
Physical data
organization
Database
CSC 240 (Blum)
6
ANSI-SPARC levels
• External
– Views: only what a user needs to see, arranged in a
convenient form
• Conceptual
– Overall logical view of database (entities, attributes,
relationships, constraints, etc.) plus some utilities
(security, integrity, etc.)
• Internal
– Specific information about where and how the data is
stored. Interfaces with operating system.
• (Physical)
– The actual stored data.
CSC 240 (Blum)
7
DBTG  ANSI-SPARC
• DBTG’s subschema corresponds to ANSISPARC’s external level (the views)
– Subschema  External schema
• DBTG’s schema is divided into two levels
in the ANSI-SPARC plan
– Conceptual schema
– Internal schema
CSC 240 (Blum)
8
Independence
• Recall that E.F. Codd’s Rules 8 and 9 called
for physical data independence and logical
data independence.
– Physical: storage changes don’t effect entities,
fields, relationships, etc.
– Logical: an extra field need not change the
views.
• ANSI-SPARC levels help provide this
independence.
CSC 240 (Blum)
9
ANSI-SPARC (Fig. 2.3 in book)
External
level
View 1
View 2
View 3
Logical data independence
Conceptual
level
Internal
level
Conceptual
Schema
Physical data independence
Internal
Schema
Physical data
organization
CSC 240 (Blum)
Database
10
Same idea/Different words
• We distinguished between prescriptive and
descriptive approaches. Other terms
include:
– Prescriptive  Procedural  (3GL)
• Step-by-step procedure for proceeding through
database record by record
– Descriptive  Non-Procedural  Declarative
 (4GL)
• Indicate what you want and let the DBMS handle it
CSC 240 (Blum)
11
4GL Tools
• Some of the standard Fourth-generation
tools include:
– Query generation
• Structured Query Language (SQL)
• Query By Example (QBE)
– Form generation
– Report generation
CSC 240 (Blum)
12
Access Objects
4GL tools in Access: Query, Form, Report and Page generators.
CSC 240 (Blum)
13
Online Help on Access Queries (gives you a
lecture)
CSC 240 (Blum)
14
Online Help on Queries
CSC 240 (Blum)
15
Online Help on Forms
CSC 240 (Blum)
16
Online Help on Access Reports
CSC 240 (Blum)
17
Data Models
• A model is “a simplified representation of a
system or phenomenon.” (Webster’s)
• A data model is a representation of the
information associated with an organization.
• When we talk about data models, we usually mean
an overall approach to representing data (defining
it, manipulating it, etc.) rather than some specific
representation of some specific organization.
CSC 240 (Blum)
18
Models and levels
• The data models to some extent reflect the level
(e.g. prescriptive vs. descriptive) that one operates
on.
– The older data models (the hierarchical and network
models) are based more in a procedural approach.
– Whereas the newer relational model is somewhat more
declarative.
– Even further from implementation details are the
Entity-Relationship and Object-Oriented models.
CSC 240 (Blum)
19
Database History: Hierarchical Model
• One of the earliest database models is the
Hierarchical Model.
– E.g. GUAM and IMS
• It is so-called because its logical structure is
hierarchical or tree-like.
• All relationships in the Hierarchical Model are of
the parent-child type.
– This is asexual reproduction, a child has one and only
one parent.
CSC 240 (Blum)
20
Example of Hierarchical Logic:
Windows Explorer
There are files in folders and folders in other folders.
CSC 240 (Blum)
21
Hierarchical (Tree-like) Structure
My Computer
A drive?
C drive
D drive
E drive
Courses
C240
C220
P201
C240wks
CSC 240 (Blum)
Web
C220wks
22
Replace the folder names with points to
obtain a graph
This kind of graph is called a tree. It has no loops.
CSC 240 (Blum)
23
Problem: what if a file could belong to
more than one folder?
• A file for CSC 240 may appear on the web page.
Does in belong in the C240 folder or the
Web\c240wks folder?
• To realize both relationships (belonging to CSC 240
and being on the web page) in the Hierarchical
Model, one must have two copies of the file.
– This would be data redundancy. And if one edits one of
the files, we could end up with an “update anomaly.”
CSC 240 (Blum)
24
Drilling down
• Another feature of the hierarchical approach
is that it requires “drilling down” (tracing
through the entire hierarchy) to get at the
data
• In the Windows Explorer example, the path
requires all of the folders
– C:Blum\Courses\C240\TheFile.txt
CSC 240 (Blum)
25
A note on file systems
• The file system (how all of the information is
stored on one’s computer) is becoming
increasingly database-like.
• The current file system typically used with
Windows XP NTFS is more like a database than
its predecessor FAT32.
• In addition Vista and Windows 7 allows the user
to opt to have files indexed (for better searching)
and also allows the user to add meta tags to file.
CSC 240 (Blum)
26
Network Model
• The Network Model arose in the early 1970s.
– The standards for the Network Model were introduced
at the Conference on Data Systems Languages
(CODASYL)
– Example of a Network Model DB: IDS
• Its logical structure is a network (a collection of
crisscrossing lines).
• Unlike the Hierarchical Model, the Network
Model’s relationships are not all of the parentchild type.
CSC 240 (Blum)
27
Example of Network Logic: A Web Site
On my web site, I have multiple links to the same set of
instructions for making graphs in Excel.
CSC 240 (Blum)
28
Network Structure
La Salle Site
My Site
My CSC 152
Other Faculty Sites
My PHY 105
XY Scatter Plot
(Depending on the connections (links), the network approach
can lessen the amount of “drilling down” needed.
CSC 240 (Blum)
29
Replace the web pages with points to
obtain a graph
This kind of graph is called a network. It has loops. The crisscrossing lines also resemble a web.
CSC 240 (Blum)
30
Relational Model: History
• Introduced by E. F. Codd (early 1970’s).
• Was an important step toward the goal of
data independence, acting on the higher
level, and all that good stuff.
• Codd dealt with the issue of redundancy
(repeated data) by introducing the concept
of normalization.
CSC 240 (Blum)
31
Relational Model: History (Cont.)
• Research versions
– System R (IBM San Jose)
• Lead to SQL
– INGRES (Berkeley)
– Peterlee Relational Test Vehicle (IBM UK)
• Early commercial versions (based on System R)
– Oracle (Oracle Corporation)
– DB2 (IBM)
CSC 240 (Blum)
32
Relational Model: Ingredients
• The main components of the Relational Model are
tables (a two-dimensional array).
• Tables are a realization of the mathematical
concept of a relation.
• Tables are reminiscent of the files used in a filebased approach.
• Table  Relation  File
• The table is logical and the data does not
necessarily take this form physically.
• A table has a name.
CSC 240 (Blum)
33
Table  Relation  File
CSC 240 (Blum)
34
Relational Model: Ingredients (Cont.)
• A table collects together associated data.
• A table is thought of in terms of rows and columns.
• The data in a single column is all of the same type, i.e.
all the same property.
– E.g. all of the people’s last names.
• The column (a.k.a. field) has a name and a type (e.g.
text, number, etc.).
• A table is distinct from a similar looking mathematical
object, the matrix, in that the order of the columns
does not matter.
• Column  Field  Attribute  Property
CSC 240 (Blum)
35
Column  Field  Attribute  Property
CSC 240 (Blum)
36
Relational Model: Ingredients (Cont.)
• The row (a.k.a. a record) collects together
the various properties that belong to a
particular object.
– E.g. a person’s first name, last name, date of
birth, etc.
• Again a table is distinct from a matrix, in
that the order of the rows does not matter.
• Row  Record  Tuple
CSC 240 (Blum)
37
Row  Record  Tuple
CSC 240 (Blum)
38
More Relational Model Vocabulary
• In addition to having a type, a field has a domain,
the set of values that the particular property is
allowed to have.
– E.g. a number must fall between 0 and 100.
– E.g. some text (string) must have two letters followed
by four numbers.
– E.g. a person’s gender must be M or F.
• Ensuring that a value falls within the domain is
called applying the domain constraint.
CSC 240 (Blum)
39
Input masks and Validation Rules are ways
to impose domain constraints in Access
CSC 240 (Blum)
40
Validation Rule example
CSC 240 (Blum)
41
More Relational Model Vocabulary (Cont.)
• The number of fields (tuples) in a table is known
as its degree.
–
–
–
–
Unary relations (1-tuples)
Binary relations (2-tuples)
Ternary relations (3-tuples)
N-ary relations (N-tuples)
• The number of records in a table is called its
cardinality.
• The degree is a property of the schema, while the
cardinality is a property of the instance.
CSC 240 (Blum)
42
cardinality
Degree and Cardinality
degree
CSC 240 (Blum)
43
Keys
• A fields or set of fields that can be used to uniquely
identify all of the rows in a table is known as a key.
• A key should not have any extraneous fields.
– E.g. if SocSecNum uniquely identifies a person, then you
don’t need SocSecNum and LastName.
• A table may have more than one field or set of
fields that serve this purpose, they are called
collectively the candidate keys.
• One key is chosen from the candidate keys to be the
primary key.
CSC 240 (Blum)
44
Keys (Cont.)
• When choosing a primary key, make sure that it must
be unique, as opposed to simply happening to be
unique for the instance you have or have in mind.
• Because of redundancy issues, it should not contain
too many fields or fields that might change.
• Be mindful of privacy issues, SocSecNum can be a
bad choice.
• For the reasons above, one often introduces an ID
field to serve as a primary field.
CSC 240 (Blum)
45
Purpose of Keys
• Keys are used to
– Uniquely identify a record as in a query
– Sort the data
– Establish relationships
• When one table’s key is found in another
table for the purpose of establishing a
relationship, it is known as a foreign key.
CSC 240 (Blum)
46
References
• Database Systems, Rob and Coronel
• http://wwwinfo.cern.ch/db/aboutdbs/classifi
cation/hierarchical.html
• Microsoft Access Help
CSC 240 (Blum)
47