Object Databases

1
Object Databases

• CC292
• Simon M. Lucas

2
Overview

• Object Graphs
• Persistence Mechanisms
• Object DBs brief history
• Lightweight Object DB db4o
• Working with db4o
• Storage
• Querying (retrieval)
• Deletion
• Transactions will be covered later

3
Object Graph ExampleNode Class

• class Node
• Node x

4
Object Graph (of Nodes)

• Node a new Node()
• Node b new Node()
• Node c new Node()
• Node d new Node()
• a.x b
• b.x d
• d.x a
• c.x d

5
Exercises

• Sketch graph of objects from previous slide
• Indicate which nodes are reachable from
• a
• b
• c
• d
• (Also indicate the shortest path distance)

6
Persistence Mechanisms

• Need to save state of an object (i.e. the values
  of its fields) for future use
• Within web application programming
• Object state is lost when the web application
  server (Tomcat) is restarted
• Therefore need some persistence mechanism
• Some approaches
• File System (either manually written code or WOX)
• Relational Database (via JDBC)
• Object Database
• Native XML Database

7
File System

• Simple to begin with for simple objects
• Can be very tedious
• Need to synchronise changes to read / write code
  and object fields (not version resilient)
• Becomes hard to manage
• Hard to query efficiently
• Still has its uses (e.g. log files)
• Use of auto XML writers such as WOX solves some
  of these problems (which ones?)

8
Relational Database

• Large installed base, well understood
• Not a good fit for objects
• Some object relational mapping needs to be done
  (tools such as Hibernate will do some of this)
• Often leads to code with string queries in
• Select from student where nameSimon
• But what if there is no matching DB field
• Errors not picked up at compile time!
• Bad news!
• Dreaded Web App Error SQL ERROR

9
Object Databases

• Potentially ideal solution
• Store objects directly in Object DB
• No special mapping process
• Save large object graphs (see below) with single
  line of code
• In practice, some interesting issues
• Referential integrity
• Transactions
• Starting to be used commercially
• Will ObjectDBs one day overtake Relational DBs?

10
Old-Style Object DB(ObjectStore, FastObjects)

• Strict adherence to principles
• Enforced referential integrity
• An object could not be deleted while other
  objects still referred to it
• Enforced transactions
• Objects could only be accessed within a
  transaction
• This effect would be cascaded (so that objects
  referenced by root object would be inaccessible
  while root was opened for access)
• Hard to work with Java (classes had to be
  modified)

11
Lightweight Object DBdb4o http//www.db4o.com

• Abandons all above principles!
• Programmer is given more responsibility
• Con possible to get things badly wrong
• Pro DB is simple to use
• Pro DB is MUCH easier to implement
• An interesting idea lets see how it works

12
Main ideas in db4o

• ObjectContainer db
• Used to save OR update Object o
• db.set(o)
• Retrieve all objects of a Given class c
• List objects db.get(c)
• List objects db.get(Product.class)
• Query by Example(QBE)
• List objects db.get(new MProduct(Sony))
• Run Queries given Predicate p
• List objects db.query(p)
• Delete an Object o
• db.delete(o)

13
Object Graphs

• db.set(o) will save all the objects reachable
  from o.
• But if one of the reachable objects changes,
  then calling db.set(o) again will not save the
  change!
• The algorithm to save an object graph is
• For each reachable object from root (o)
• Save if not already saved

14
Product Example

• A Product has a name, and a price
• A Manufactured Product also has a Manufacturer
• A manufacturer has a name and a URL
• In this example well see how to save objects of
  these types
• And how to write queries for them

15
The Data Classes

• public class Product
• String name
• double price
• // constructor and toString() omitted
• public class MProduct extends Product
• Manufacturer manufacturer
• public class Manufacturer
• String name
• String url

16
Saving Some Objects

• public static void main(String args) throws
  Exception
• ObjectContainer db Db4o.openFile(dbName)
  
• db.set(new Product("USB Splash 1 Gb",
  29.99))
• db.set(new Product("USB Flash 256 Mb",
  19.99))
• Manufacturer sony new
  Manufacturer("Sony", "www.sony.com")
• MProduct mp new MProduct("USB
  Flash 256 Mb", 22.99, sony)
• db.set(mp)
• sony.url "http//apple.com"
• db.set(new MProduct( "USB Flash
  Drive 8 Mb", 2.99, sony))
• System.out.println("Set the objects")
• db.close()

17
Running Queries

• There are several ways to run queries
• (Query By Example (QBE), SODA, and Predicates)
• Here well just cover the Predicate method
• Main concept
• Define a Predicate that takes objects of a
  certain class as an argument
• The Predicate should return true if the object
  should be returned, false otherwise

18
Example Predicates

• Two examples
• Our Manufacturer predicate always returns true
• Hence all the Manufacturers in the DB are
  returned
• Our Product predicate selects all products whose
  name contains a given substring

19
All Manufacturers Query(how could you simplify
this?)

• public static void manufacturers(ObjectContainer
  db)
• ListltManufacturergt set db.query(new
  PredicateltManufacturergt()
• public boolean match(Manufacturer m)
• return true
• )
• // now do something with the result
• for (Manufacturer m set)
• System.out.println(m)

20
Like Products Query

• public static void likeProductQuery(ObjectContain
  er db, final String sub)
• ListltProductgt set db.query(new
• PredicateltProductgt()
• public boolean match(Product p) //
  this tests of sub is a // sub-string of
  p.name
• return p.name.indexOf(sub) ! -1
• )
• for (Product p set)
• System.out.println(p)

21
Running The Queries

• public static void main(String args)
• ObjectContainer db
• Db4o.openFile(SaveProducts.dbName)
• // print all products with Flash in title
• likeProductQuery(db, "Flash")
• System.out.println("")
• // print all manufacturers
• manufacturers(db)
• db.close()

22
Sample Output

• Exercise Fill in what the output would be!

23
Retrieval Depth

• An important concept
• Given an object graph, compute the shortest path
  to every reachable object
• When retrieving an Object
• only those objects within depth distance from the
  root object are retrieved
• All other references set to null
• The root object in this case, is the argument
  to the predicate
• Default depth in db4o is 5
• Exercise underline the root object in each of
  the previous queries.

24
Summary

• So far weve seen how easy a lightweight Object
  DB such as db4o can be to use
• Some unresolved issues
• Query efficiency in our example the predicate
  had to be tested against every instance of an
  appropriate class!
• Transactions to be covered later
• Future prospects very exciting
• Will use db4o in lab and for the assignments