Designing OQL: Allowing Objects to be Queried

1
Designing OQL Allowing Objects to be Queried

• Sophie Cluet
• INRIA
• Presented By
• Ying Liu
• April 9th, 2003

2
ODMG OQL User Manual

• Besides the paper, I referred the ODMG OQL User
  Manual.
• http//www.cis.upenn.edu/cis550/oql.pdf

3
History

• In 1988, INRIA in Altair Group started to develop
  an OODBS
• In 1991, O2 DBS as a product of O2 Technology,
  and O2SQL as the query language
• In 1991, ODMG was created and O2SQl became the
  standard of OQL
• In 1998, OQL 5.0.

4
O2 System Overview(http//www.cis.upenn.edu/cis5
50/oql.pdf)
5
ODMG Standard

• 1. an object model
• 2. an object definition language for this model,
  with its own syntax, ODL
• or its expression through C and Smalltalk
  syntax
• 3. an object query language for this model, OQL
• 4. a C binding allowing C programs to operate
  on a database compliant to the object model
• 5. a Smalltalk binding allowing Smalltalk
  programs to operate on a database compliant to
  the object model

6
ODMG Model

• The ODMG object model supports the notion of
  classes, of objects with attributes and methods,
  of inheritance and specialization.
• It offers the classical types to deal with
  string, date, time, time interval and timestamp.
• It supports the notions of relationships and
  collections .

7
Basic Constructs

• Atomic Integer, Float, Boolean, Char
• Pre-built String, Bitmap, Date
• Collection
• Set unordered, duplicate free collection of
  arbitrary cardinality
• List Ordered collection of arbitrary
  cardinality
• Bag Unordered collection of arbitrary
  cardinality
• Array Ordered collection of fixed cardinality.

8
An O2 Schema Example
9
Entry points

• To query any database, we need various entry
  points.
• In O2 these are the named objects and named
  values.
• For example, Globe is an entry point.

10
Creating an object

• We create objects using struct, list, array, bag
  and set.
• This creates a temporary object with the name
  attribute initialized to Trent.
• An object collection can be created in the same
  way.

11
Simple Queries

• Atomic Values

Struct Values
12
SelectFromWhere

• Set (a set is an un-ordered, duplicate-free
  collection)
• OQL uses the in keyword to define the range of
  a variable.

13
SelectFromWhere (contd)

• List or Array A list or an array is an ordered
  collection that can contain duplicate elements.
• Since it is ordered, you may extract any of its
  elements if you know their position.
• Join

14
Constructing results

• We can also construct a query result with an
  explicit structure using the struct, set, bag,
  list and array constructors.

15
Constructing results (contd)

• Another example

16
Operators

• Count
• Define We can name the result of a query using
  the define clause. For example,

17
Operators (contd)

• Exits
• Group By

18
Operators (contd)

• Like
• Order by The result of an order by operation is
  always a list

19
Set operators

• Union (), Interaction () and Difference (-)

20
Conversions

• List to set
• Set to list
• Flatten

21
Combining operators

• every operator can be combined with any other
  operator.

22
Indexes

• Defining an index

23
OQL More

• The most important characteristics of the OO
  language and system include
• Hierarchy and inheritance
• Method encapsulation
• Polymorphism

24
A schema
25
More query

• Path expressions

26
Method invoking

• Method invoking

27
Polymorphism

• Late binding
• Class indicator

28
Summary

• Basic queries
• To query any database you need various entry
  points. In O2 these are the named objects and
  named values.
• Select..from..where
• Constructing results
• The structure of a query result is very often
  implicit . However , you can also construct a
  query result with an explicit structure using the
  struct, set and list constructors.
• Operators
• OQL operators include define, order by, count,
  exists, group by and like. They can be combined
  for complex queries.
• Indexes
• Method Invoking
• Polymorphism
• Path expression