Mapping Java Objects to XML and Relational...

Colorado Software Summit: October 24 – 29, 2004 © Copyright 2004, Oracle Corporation

Donald Smith — Mapping Java Objects to XML and Relational Databases: O-R and O-X Mapping

Mapping Java Objects to XML and Relational DatabasesO-R and O-X Mapping

Donald SmithTechnology Director, Oracle Corporationblog: jroller.com/page/dsmith

http://www.jroller.com/page/dsmith



Speaker’s Qualifications

Decade of experience in OO PersistencePresented at Java One, Oracle World, OOPSLA, JAOO, Sun Tech Days, TheServerSide Symposium, etc.Author of numerous articles on persistence challenges



About the Audience…

Who considers themselves first and foremost to be a DBA or “Database expert”?Who considers themselves first and foremost to be a Java and/or Web Services developer?Who considers themselves first and foremost to be an Architect?Who considers themselves first and foremost to be a manager, and will you admit it?



Goal

Discuss and demonstrate the challenges of mapping Java Objects to Relational Databases and XML documents



AgendaQuick Review of Java Mapping

O-R and O-X OverviewTerminology

O-R MappingDirect, 1-1, Aggregate, 1-M, M-M, Inheritance

O-X MappingDirect, Composite, Transformation, Inheritance





O-R MappingO-X Mapping



Impedance Mismatch

Difference in relational, XML and object technology known as “impedance mismatch”Challenging problem

Requires relational, XML and object expertise



O-R Impedance MismatchFactor J2EE Relational DatabaseLogical Data Format

Objects, methods, inheritance

Tables, SQL, stored procedures

Scale Hundreds of megs Gigabytes, terabytes

Relationship Memory references Foreign keysUniqueness Internal object identity Primary keysKey Skills Java development,

object modelingSQL, Stored Procedures, data management, Data modeling

Tools IDE, Source code management, Object Modeler

Schema designer, query manager, performance profilers, database config

Corporate Org. Structure

“Newer technology”often with weak organizational ties to database mgmt

Often mature infrastructure with significant legacy considerations

Tech

nica

lB

usin

ess

Pol

itica

l



O-X Impedance MismatchFactor J2EE XMLLogical Data Format

Objects, methods, inheritance

XSD, XPATH

Scale Hundreds of megs Depends

Relationship Memory references Generally use AggregationUniqueness Internal object identity Unique IdentifierKey Skills Java development,

object modelingXML, XSD, XPATH, XQUERY

Tools IDE, Source code management, Object Modeler

XML Design Tools and Viewers

Corporate Org. Structure

Often integrating or using legacy application code

High expectations of data portability

Tech

nica

lB

usin

ess

Pol

itica

l



JDBC

Java standard for accessing data sourcesSimply – data source connectionOK with:

“Window on data” applicationsBusiness logic entrenched on databaseJava nothing more than GUI tool

SQLrows

JDBC

SQLrows

Java & Java & Web Web

ServicesServices



Persistence LayerAbstract persistence details from application layerRequires Object to Data source mapping…

SQLrows

Objects

Persistence Layer

Objectsobject-level

querying and creationresults are objects

results arereturned as

raw data

API uses SQLor databasespecific calls

object creation and updates through object-level API

Java & Java & Web Web

ServicesServices



O-R Mapping Tools

O-R Mapping Tools help you manage the mapping’s meta data



3 Levels of XML Representation

XML ParserO-X Data Converter

XMLDocument

UnmanagedObject

O-X Persistence Manager

XDB EIS

JDBC J2C

DOM ManagedObject/EJB

File

WebService BPM JMS RDB



XML ParserJAXP – Java API for XML Parsing

DOMSAX

Very low levelAkin to straight using straight JDBC for database interactionsUseful for simple and raw GUI based apps where a business model is overkill

XML Parser



O-X Data ConverterConverts XML data to Java objects and vice-versaAccessed by applications through Marshal/Unmarshal interfaceUsually for non-transactional data sourcesJAXB implementations

O-X Data Converter



O-X Data Converter Limitations

Beware of your JAXB implementation…Most generate Java classes from XML Schema –static, inflexibleMost offer no control over the mappingsCan’t use your own Java classesApplication code is tightly coupled to a specific XSD

Usually no GUI tools to do mappingsConversion only, no run-time manager available for transactional data sourcesHomogeneous data support

Specific interfaces and generation for XML



O-X Persistence ManagerFlexible mapping, developers control how objects are mapped to XML – “meet in the middle”

Can use developer-defined Java classesIndependence between object model and XML schemaBusiness logic can be safely added into Java modelClasses can be mapped to multiple schemas – vice versaCan be JAXB compliant





Support complex XML mappingsPositional, path informationExamples coming…

May provide visual mapping interfaceMay support other data sources – relational and EIS




O-X Persistence ManagerPersistence manager functionality may be required for transactional XML data sources such as EIS systems, XML databases.Provides additional capabilities on top of data conversion such as:

CachingQueryingTransactionsConcurrency



O-X Mapping

Map Object Model to XSDEither code gen XSD from Object Model, vice versa, or “meet in the middle”

Java IDEJava source

MappingTool

XSD

MappingXML



O-X Mapping Tools

O-X Mapping Tools help you manage the mapping’s meta data



O-R versus O-XO-X

Much more aggregation (aka “denormalized”)Relationships represented in structure of documentRelated data usually contained in single “bite”• “Less Chatty”

O-RRelationships represented by data in table (FK)Frequent round-trippingData storage and practices more standardized



Mapping TerminologyDirect Mappings

Field in Object model maps to field in datasource• May require conversion, aggregation or serialization

Relationship MappingsThought of always as uni-directional• “Bi-directional” relationships are two distinct

mappings

Source is the originator of the relationshipTarget is referenced object or set of objects from the source



Mapping Terminologypublic class Source {private String directAttribute;private Target target;private Collection targetCollection;

}

Source

String:directAttribute

Targettarget

targetCollection

*



AgendaQuick Review of Java MappingO-R Mapping

Direct, 1-1, Aggregate, 1-M, M-M, Inheritance

O-X Mapping



Simple Direct Mapping

Customer

id: intname: StringcreditRating: int

CUSTID NAME C_RATING



Direct Conversion Mappings

Customer

id: intname: Stringgender: String

CUST

ID NAME GENDERFFemaleMMaleRO

Employeeid: intname: StringworkTerm: Date[]

EMPLOYEE

ID NAME START END



1-1 Object ModelCustomer


Address

id: intcity: Stringzip: String

1:1 Relationship



Typical 1-1 Relationship

CUST

ID NAME A_IDC_RATINGADDR

ID CITY ZIP

CUSTID NAME A_ID

ADDRID CITY ZIP

7 Don 3 3 Ottawa K2J4X717 Toronto K1P2T213 Trisha 17



1-1 FK in Target Table

CUST

ID NAME C_RATE C_ID

ADDR

ID CITY ZIP

CUSTID NAME C_ID

ADDRID CITY ZIP

7 Don 73 Ottawa K2J4X717 Toronto K1P2T213 Trisha 13



Association TableCUST

ID NAME C_RATING

ADDR

ID CITY ZIP

A_ID

CUST_ADDR

C_ID

CUSTID NAME

C_ID A_ID

ADDRID CITY ZIP

7 Don7 3

3 Ottawa K2J4X717 Toronto K1P2T213 Trisha

13 17

CUST_ADDR



Aggregation

CUST

ID NAME CITY ZIPC_RATING

CUSTID NAME CITY ZIP7 Don Ottawa K2J4X7

Toronto K1P2T213 Trisha



Multiple Table Mappings

CUST

ID NAME A_ID

ADDR

ID CITY ZIP

CUST_CREDIT

ID C_RATING

CUST

ID NAME

ADDR

ID CITY ZIP

CUST_CREDIT

ID C_RATING A_ID




CUST

ID NAME

CUST_CREDIT

ID C_RATING

ADDR

ID CITY ZIP C_ID

CUSTID NAME

CUST_CREDITID C_RATING

ADDRID CITY ZIP C_ID




CUST

ID NAME

ADDR

ID CITY ZIP

CUST_CREDIT

ID C_RATING A_IDCC_ID



1-M Object ModelCustomer


Phoneid: intString: typeString: areaCodeString: number

1:M Relationship*



1-M Relationship

CUST

ID NAME C_RATING PHONE_NUMBER

ID TYPE A_CODE PNUMB C_ID



1-M Relationship Trick

Mapping a 1-1 “back reference” from the target of the 1-M to the sourceAllows persistence layer to more easily handle the foreign key

Customer



*

1



1-M Relationship Trap!What if order is important?Specify field to order by

Customer



*



1-M Relationship Trap!What if collection indexis the order criteria?

Customer



*

:Customereename = “Don”

:Collectionee

:Phone1eetype = ‘home’numb = 45367

:PhoneNeetype = ‘cell’numb = 6573

…

1 n



1-M Relationship Trap!Challenging problem because “state” from Java utility class needs to be persistedBest solution is to find or make a persistent attribute of the target the order by fieldRegardless, database requires SOMETHING to order by



1-M Relationship Trap!

:Employeeeename = “Don”

:Collectionee

:Phone1eetype = ‘home’numb = 45367

:Phone1eetype = ‘cell’numb = 6573

…

1 n

EMPID NAME

PHONEID TYPE NUMBER

7 Don

3 home 4536717 cell 6573

ORDER13

15 work 99882 2



M-M Object ModelEmployee

id: intname: String

Project

id: intname: Stringbudget: double

M:M Relationship *

*



M-M RelationshipEMP

ID NAME

PROJECT

ID NAME BUDGET

P_ID

EMP_PROJ

E_ID

EMPID NAME

E_ID P_ID

PROJECTID NAME BUDGET

7 Don7 3

3 Tiger 5,000,00017 Dragon 1,000,00013 Mike

7 17

EMP_PROJ



M-M Relationship Trap!

EMP

ID NAME

PROJECT

ID NAME BUDGET

P_ID

EMP_PROJ

E_ID ROLE

Common challenge – RDBs often intertwine data and relationships. Where would this go in object model?



M-M Relationship Trap!In relational world, putting “data” on a relationship is not uncommonNot possible in JavaMust store data in an object – requires some re-thinking of Object model



M-M Trap SolutionEmployee

id: intname: String

Project

id: intname: Stringbudget: double

*

*

Responsibility

roleName: Stringemp: Employeeproj: Project



Direct Collection Mapping

:Employeeeename = “Don”tasks = {“schmooze”, “gossip”, “coffee”}

EMPID NAME

EMP_TASKID TASK

7 Don

7 schmooze

7 gossip

7 coffee



Direct Collection IssuesInefficient for large direct collections

Changes require lots of “change detection” analysis, or massive purge and insert on database

Only useful for PRIMITIVES, more than 1 piece of date requires 1-M mappingConsider Maps of Primitives…



Direct Map Mapping

:Employeeeename = “Don”contact = {“email” -> “[email protected]”, “home” -> “555-1212”,“cell” -> “555-2323”}

EMPID NAME

CONTACTID KEY

7 Don

7 email

7 home

7 cell

VALUE

[email protected]

555-1212

555-2323



InheritanceVery straight forward at first, but can become challenging because of the number of possible solutionsCompounded with relationshipsCan be source of contention with DBA –generally leads to lots of joining on queries!



“Leaf Table Mapping”

Sub1sub1var: String

Root {a}

id: intvar: String

Sub2 {a}sub2var: int

Sub22sub22var: String

Map only concrete classes

SUB1ID VAR

SUB22

7 FooSUB1VAR

Pish

ID VAR13 Tim

SUB2VAR677

SUB22VARFlim



“Leaf Table” IssuesHow do you query for the Root class?Data is not normalized



“All Table Mapping”

Sub1sub1var: String

Root {a}

id: intvar: String



Map all classes ROOT

ID VAR

SUB2

7 Foo13 Bar

ID13

SUB2VAR677

SUB22VARFlim

SUB1ID7

SUB1VARPish

ID13

SUB22

Problem! How to tell type?!



Identifying Type

Sub1sub1var: String

Root {a}

id: intvar: String



Map all classes ROOT

ID VAR

SUB2

7 Foo13 Bar

ID13

SUB2VAR677

SUB22VARFlim

SUB1ID7

SUB1VARPish

ID13

SUB22

Need Type Identifier!

TYPES1

S22



Identifying Type

ROOTID VAR7 Foo

13 Bar

Lots of strategies exist…TYPE

S1S22 ROOT

ID VAR7 Foo13 Bar

TYPEcom.foo.Sub1

com.foo.Sub22

ROOTID VAR7 Foo

13 Bar

if id < 10 type = Sub1

else type = Sub22



“All Table” IssuesLots of joiningType identification can be inefficient



“Single Table Mapping”

Sub1sub1var: String

Root {a}

id: intvar: String



Map to one table

ROOTID VAR7 Foo

13 Bar

SUB1VARPishnull

SUB2VARnull677

SUB22VARnullFlim



Identifying Type

Sub1sub1var: String

Root {a}

id: intvar: String



ROOTID VAR7 Foo13 Bar

SUB1VARPishnull

SUB2VARnull677

SUB22VARnullFlim

TYPES1

SUB22

if SUB1VAR != null type = Sub1

else type = Sub22

Two options for type detection



“Single Table” IssuesDBA will freak out – de-normalized databaseType identification can be inefficient



“Combination Table Mapping”

Sub1sub1var: String

Root {a}

id: intvar: String



Map appropriate classes

ROOTID VAR7 Foo13 Bar

SUB2VAR677

SUB22VARFlim

SUB1ID7

SUB1VARPish

ID13

SUB22

TYPES1

S22



Which Strategy to Use?Leaf Table

When never querying at abstract class levelAll Table

When not doing much querying• Too many joins

Single TableWhen one subclass is much more prevalent, and your DBA doesn’t care about normalization

CombinationUse to optimize above situations



Wouldn’t It Be Cool?

Sub1sub1var: String

Root {a}

id: intvar: String



Map to view for reads…ROOT (View)ID VAR7 Foo13 Bar

SUB1VARPishnull

SUB2VARnull677

SUB22VARnullFlim

ROOTID VAR SUB27 Foo13 Bar ID

13SUB2VAR

677

SUB22VARFlim

SUB1ID7

SUB1VARPish ID

13

SUB22

Map to tables for writes…



AgendaQuick Review of Java MappingO-R MappingO-X Mapping

Direct, Composite, Transformation, Inheritance



Example Object ModelAddress

addressee: Stringcity: Stringstreet1: Stringstreet2: Stringstate: Stringcountry: StringzipCode: String

LineItemlineNumber: longitemName: Stringquantity: longitemPrice: BigDecimal

Order

id: longorderedBy: String address

1

* lineItems



Direct Mapping

<ORDER ORDER_ID="1234"><ORDERED_BY>Jane Doe</ORDERED_BY><ADDRESS>

...</ADDRESS><LINES>

...</LINES><LINES>

...</LINES>

</ORDER>

: Order

id = 1234orderedBy = “Jane Doe”

: Address

: LineItem : LineItem



Composite Object Mapping



...</LINES><LINES>

...</LINES>

</ORDER>

: Order


: Address




Composite Collection Mapping



...</LINES><LINES>

...</LINES>

</ORDER>

: Order


: Address


Possible <lineitems> tag here



Positional InformationGood O-X Support

JAXB/Class Generation Today…

<ADDRESS><ADDRESSEE>Jane Doe</ADDRESSEE><STREET>123 Some St.</STREET><STREET>Apt. #123</STREET>

…</ADDRESS>

: Address

addressee = “Jane Doe”street = “123 Some St.”apartment = “Apt. #123”…

<ADDRESS><ADDRESSEE>Jane Doe</ADDRESSEE><STREET>123 Some St.</STREET><STREET>Apt. #123</STREET>

…</ADDRESS>

: Address

addressee = “Jane Doe”street = {“123 Some St.”,

“Apt. #123”}…



Path InformationGood O-X Mapping: LineItem

lineNumber = 1itemName = “Pens”itemPrice = 2.50quantity = 50

<LINES LINE_NO="1"><ITEM>

<NAME>Pens</NAME><PRICE>2.50</PRICE>

</ITEM><QUANTITY>50</QUANTITY>

</LINES>

<LINES LINE_NO="1"><ITEM>

<NAME>Pens</NAME><PRICE>2.50</PRICE>

</ITEM><QUANTITY>50</QUANTITY>

</LINES>

: Lines

lineNumber = 1quantity = 50

: Item

name = “Pens”price = 2.50

JAXB/Class Generation Today…



Direct Collection Mapping

<EMPLOYEE><TASK>this</TASK><TASK>that</TASK><TASK>other thing</TASK>

</EMPLOYEE>

: Employee

tasks = {“this”, “that”, “other thing”}

• Compare with Positional Mapping



Transformation Mapping

<EMPLOYEE><START-TIME>9:00:00</START-TIME><END-TIME>17:00:00</END-TIME>

</EMPLOYEE>

: EmployeenormalHours

= {9am,5pm}

: FieldTransformer

: FieldTransformer

Marshal (Write)

: AttributeTransformer

<EMPLOYEE><START-TIME>9:00:00</START-TIME><END-TIME>17:00:00</END-TIME>

</EMPLOYEE>

: EmployeenormalHours

= {9am,5pm}

Unmarshal (Read)



Conversion Mapping

<EMPLOYEE><GENDER>F</GENDER>

</EMPLOYEE>

: Employee

gender = “Female” “Female” to “F”“Male” to “M”



Simple Type Translator

<element name=“AREA-CODE” type=“anySimpleType”/><element name=“NUMBER” type=“anySimpleType”/>

PhoneNumberareaCode: Objectnumber: Object

Class Model

<PHONE-NUMBER><AREA-CODE xsi:type=“xsd:int”>613</AREA-CODE><NUMBER xsi:type=“xsd:string”>KL51234</NUMBER>

</PHONE-NUMBER>

: PhoneNumberareaCode = 613number = “KL51234”

Instance Model



Namespace Support

<abc:EMPLOYEE xmlns:abc=“http://X” xmlns:def=“http://Y”><def:FIRST-NAME>Jane</def:FIRST-NAME><def:LAST-NAME>Doe</def:LAST-NAME>

</abc:EMPLOYEE>

<ns1:EMPLOYEE xmlns:ns1=“http://X” xmlns:ns2=“http://Y”><ns2:FIRST-NAME>Jane</ns2:FIRST-NAME><ns2:LAST-NAME>Doe</ns2:LAST-NAME>

</ns1:EMPLOYEE>

: Employee

firstName = “Jane”lastName = “Doe”

Employee Default Root = abc:EMPLOYEEXPath firstName = def:FIRST-NAMEXPath lastName = def:LAST-NAME

Namespace Resolver:abc = http://Xdef = http://Y



Inheritance<element name=“ADDRESS” type=“ADDRESS-TYPE”/><complexType name=“ADDRESS-TYPE”>

…..</complexType><complexType name=“CANADIAN-ADDRESS-TYPE>

<complexContent><extension base=“ADDRESS-TYPE”>

…</extension>

</complexContent></complexType>

Address

street: Stringcity: String

CanadianAddress

province: StringpostalCode: String

ClassModel

: CanadianAddress

street = “123 Any St.”city = “Ottawa”province = “ON”postalCode = “A1B 2C3”

<ADDRESS xsi:type=“CANADIAN-ADDRESS-TYPE”><STREET>123 Any St.</STREET><CITY>Ottawa</CITY><PROVINCE>ON</PROVINCE><POSTAL-CODE>A1B 2C3</POSTAL-CODE>

</EMPLOYEE>

InstanceModel



Conclusions

Much more to O-R and O-X Mapping than first meets the eyeO-R and O-X Share several common challenges, but each have their own unique perspectives



Recommended Sessions…Donald Smith

Impact of Tiger on Java Persistence

Mike KeithStandards Based Persistence through Web ServicesEvolution of the EJB Entity

Neil GrahamLower-level Data Manipulation with J2SE 5.0

Matthew Wakeling[Advanced Techniques | Performance] in Object Warehousing

Mapping Java Objects to XML and Relational...

Documents

Transcript of Mapping Java Objects to XML and Relational...