You are on page 1of 45

Oracle Trading

Community Architecture
Understanding the Concepts of
Data Quality Management
An Oracle White Paper
December 2004

Understanding the Concepts of


Data Quality Management

Table of Contents
Executive Overview..................................................4
Introduction................................................................5
The Need for Clean TCA Data....................................5
Data Quality management (DQM)........................6
Overview.....................................................................6
How DQM Works........................................................6
Transformation Functions..............................................7
Word Replacement Lists................................................8
Match Rules..............................................................10
Match Rule Types.......................................................10
Defining Match Rules..................................................11
DQM In the E-Business Suite and Customer Hub
Solution.....................................................................17
Smart Search............................................................17
Duplicate Prevention...................................................17
Duplicate Identification...............................................17
Conclusion................................................................19
Appendix A................................................................20
DQM Examples.........................................................20
Smart Search Example................................................20
Duplicate Prevention Example......................................25
Duplicate Identification Example...................................31
What To Do With Duplicates.....................................33
Appendix B................................................................34
Definitions and Terminology.....................................34
Appendix C................................................................36
E-Business Suite Applications That Have Registered Merge
Routines with Party Merge..............................................36

Appendix D................................................................38
E-Business Suite Applications Referenced by Customer Account
Merge.........................................................................38

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 2

Understanding the Concepts of


Data Quality Management

Executive Overview
This document is intended to provide the reader with an
overview of the functionality and usage of the Oracle Data
Quality Management (DQM) engine, which is designed to help
a deploying company identify duplicate records within a
customer registry.
The primary audience for this document includes individuals
within a deploying company who are responsible for
information quality and consultants who are implementing
DQM on behalf of a deploying company.
This white paper serves as a high level guide for understanding
DQM and how it is used. For more information regarding DQM
refer to the following documents available on OracleMetaLink:

Oracle Trading Community Architecture Administration


& Reference Guides

Oracle Data Librarian User & Implementation Guides

Please refer to Appendix B at the end of this paper for a list of


Definitions and Terminology mentioned throughout the
document.

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 3

Introduction
A
trading
community
consists
of
the
organizations, people and
places
with
which
a
company does or intends to
do
Thebusiness.
Oracle Customer Data
Hub is a packaged solution
that allows a deploying
company to create a single,
enterprise-wide
customer
database by consolidating
customer
information
residing
in
different
heterogeneous systems.

The Data Quality Management (DQM) tool is a flexible


duplicate identification engine that is bundled with the Oracle
Trading Community Architecture (TCA). TCA stores
information and relationships about a companys trading
community, and is the core data model used by all E-Business
Suite applications as well as the Oracle Customer Data Hub
solution.
The features of the DQM engine can be used with other TCA
data quality tools to not only identify duplicate records, but
also resolve them by merging the duplicate parties and
accounts while properly maintaining all associated
transactions.

The Need for Clean TCA Data


An accurate representation of the trading community is
essential to the success of any company. With duplicated and
fragmented data, companies are susceptible to poor customer
interactions, low customer satisfaction rates and low customer
retention rates. In todays increasingly competitive business
environment, gaining, and more importantly retaining
customers is priority number one.

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 4

Data Quality management (DQM)


Overview
DQM is a highly configurable duplicate identification engine.
The flexibility of the DQM engine allows a deploying company
to create their own definitions of what constitutes a duplicate
record. Alternate definitions can be used to support specific
business functions. These definitions are called Match Rules,
which can be leveraged to:
-

enhance TCA registry search screens

prevent duplicate records from entering the TCA


registry

identify existing duplicate records within the TCA


registry

How DQM Works


At the core of the DQM search engine is the ability to identify
similar records where the data is misspelled or inconsistent.
Many traditional search engines only have the ability to
identify exact records. For example, imagine that the
following parties exist in a TCA registry:

The Input Value is the


value that the DQM engine
uses to search the TCA
registry to find matching
records.

Business World

Business Wurld

Bisness World

Using a traditional searching method a user tries to find a


record called Business World, which becomes the input
value:
TCA Registry

Input Value
Business World

busines world
Business Wurld

Matching
Records
None

Bisness World

Figure 1: Traditional Search Engine

As you can see, no records were returned from this search


because none of the records in the TCA registry matched the

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 5

input value exactly. However the data in the TCA registry


shows that there are other Business Worlds, which happen to
be misspelled. The DQM engine was designed to find these
misspelled records by searching for similar records instead
of exact records. To do this DQM uses Transformation
Functions.
Transformation Functions

Staging
Tables
store
transformed TCA registry
data solely for the purpose
of
identifying
duplicate
records.

A Routine is a procedure
that transforms a data
element in a certain way.
Routines are called upon by
Transformation Functions.

Transformation Functions take data stored within specified


attributes of the TCA registry and alters it for the purpose of
identifying duplicates. The transformed data is stored in
Staging Tables and is solely used internally by the DQM engine
to identify matching records. Transformed data is helpful
when trying to identify similar records with words that are
aliased, inconsistent or misspelled. DQM comes with several
seeded transformation functions, plus the flexibility to create
custom transformation functions to address unique
requirements.
Two of the basic seeded transformation functions, EXACT
and CLEANSE, are listed below. The Routines column
shows the routines that are called to transform the data. A
transformation function can contain one to many routines. For
a complete list of all seeded transformation functions please
refer to the document Oracle Trading Community
Architecture Reference Guide available on OracleMetaLink.
Transformation
Function
EXACT

Routines
Capitalize all letters
Remove non-alphanumeric
characters
Reduce all white space to a single
white space

CLEANSE

Capitalize all letters


Remove non-alphanumeric
characters
Reduce all white space to a single
white space
Remove double letters
Remove all vowels, except leading
vowels

Remember in Figure 1 on Page 5, no matches were found when


searching for Business World because the search engine
used could only identify exact matches. Now lets see how
the DQM search engine would handle this same scenario using

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 6

transformation functions to find similar matches instead of


exact matches.
In Figure 2 below, notice how the TCA registry data is
transformed and placed into a staging table for the purpose of
identifying duplicates. In this example the transformation
functions used are EXACT and CLEANSE.
First, the input value is transformed. Then it is evaluated
against the data in the staging table to look for duplicate
records. Because the data has been transformed, the DQM
search engine finds three matching records and returns them
to the user.

TCA Registry

Input Value
Business World

busines world
Business Wurld
Bisness World

TCA Registry
Matching Records
busines world
Business Wurld
Bisness World

Staging Table

Transformed
Input Value
BSNS WRLD

BSNS WRLD
BSNS WRLD
BSNS WRLD

Staging Table
Matching Records
BSNS WRLD
BSNS WRLD
BSNS WRLD

Figure 2: DQM Search Engine

Figure 2 shows that by using the DQM search engine, the


common misspellings of Business World were neutralized by
applying two basic transformation functions.
It is important to remember that the actual data in the TCA
Registry is not being physically changed by the transformation
functions. It is only being copied and then transformed into a

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 7

temporary Staging Table for the purpose of identifying


duplicate records.
In addition to calling routines, transformation functions also
can call Word Replacement Lists to help alter and standardize
the data.
Word Replacement Lists

Word Replacement Lists (WRL) standardize words or remove


noise words to improve the likelihood of finding duplicates.
They are especially helpful with nicknames and abbreviations.
Similar to routines, WRLs can be called upon by transformation
functions.
DQM comes seeded with a number of Word Replacement Lists
relating to street types, US state names, person names, and
organization names. Below are samples from the seeded
Person Name and Organization Name Word Replacement
Lists to demonstrate how words can be standardized for the
purpose of matching.

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 8

Person Name WRL


TCA
Standardiz
Bill
Bill

Organization Name WRL


TCA
Standardize
Inc
Inc.

Billy

Bill

Incorp.

Inc.

William

Bill

Incorporated

Inc.

To give an example of how WRLs can be used to identify


duplicate records, Figure 3 below shows the application of the
Organization Name WRL.
First, the input value is transformed using a WRL. Then it is
evaluated against the data in the staging table to look for
matching records. Because the data has been transformed
using the WRL, the DQM engine finds three duplicate records
and returns them to the user.

TCA Registry

Input Value
Business Incorp

Business Inc
Business Incorp.
Business Incorporated

TCA Registry
Matching Records
Business Inc
Business Incorp.
Business Incorporated

Staging Table

Transformed
Input Value
Business Inc.

Business Inc.
Business Inc.
Business Inc.

Staging Table
Matching Records
Business Inc.
Business Inc.
Business Inc.

Figure 3: Using Word Replacement Lists to identify


duplicate records

Seeded Word Replacement Lists are provided for American


English and can be updated or new lists can be created.
Additionally, deploying companies can use the existing
architecture to create word replacement lists for any language

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 9

around the world. Both single and double byte character sets
are supported.
Using Transformation Functions and Word Replacement Lists
to alter Attributes for the purpose of finding matching records
is how the DQM engine works. However, a key differentiator
of the DQM tool is that the user has the flexibility to define
what they consider to be a duplicate. To do this, selected
Attributes, Transformation Functions and Word Replacement
Lists are packaged together to create Match Rules.

Match Rules
Match Rules represent the different definitions of what
constitutes a duplicate record, and are used by the DQM
engine to identify similar records in the TCA registry. A
deploying company can create any number of match rules to be
used at different times for different purposes. A number of
match rules are seeded to provide a foundation for using the
DQM engine, however users can additionally choose to create
their own match rules based on specific requirements.
Match Rule Types

There are three types of match rules: Search, Bulk and


Expanded.
Search Match Rules
Search Match Rules are used by DQM to enhance party
search screens or to prevent duplicate records from
entering the TCA registry through party creation screens.
These match rules take user data input and search the TCA
registry for potential duplicates.
interMedia Indexes are
built using inverted indexes,
a structure that differs from
normal indexes and allows
the use of a contains
query.

Search Match Rules are comprehensive in their search


because they use interMedia Indexing, which is a type of
indexing that supports the use of a contains query. This
allows Business World to be found as a potential duplicate
for World of Business.
Bulk Match Rules
Bulk Match Rules are mainly used by DQM to prevent
duplicate records from entering the TCA registry through
bulk data import, but can also be used to identify existing
duplicates within the TCA registry. When used to prevent
duplicates from entering through bulk import, these match
rules take each imported record one at a time and search
the TCA registry for potential duplicates.
Bulk Match Rules are less comprehensive than Expanded
Match Rules but provide improved performance. This is

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 10

because they use B-Tree Indexing rather than interMedia


Indexing.
Expanded Match Rules
Expanded Match Rules are used by DQM to identify existing
duplicate records within the TCA registry. These match
rules take an existing record as the input and search the
remainder of the TCA registry looking for potential
duplicates.
Expanded Match Rules are as comprehensive as Search
Match Rules because they also use interMedia Indexing.

Defining Match Rules

The Work Unit consists of


all records from the Staging
Tables with attribute values
that match the transformed
acquisition attribute values
of the input record.
The
work unit saves time and
resources
because
this
relevant subset of records,
not
the
entire
staging
tables, can be compared
against the input record for
scoring.

To effectively define match rules it is helpful to understand how


they are used. The DQM engine uses match rules in a two-step
process consisting of an Acquisition phase and a Scoring
phase. The Acquisition phase takes an input value and
searches the TCA registry for similar records. These
matching records form a subset called the Work Unit. The
Scoring phase then evaluates the records in the Work Unit to
determine which records are closer matches to the input value
than others. Match rules provide the DQM engine with the
information necessary to execute both of these phases.
Figure 4 shows the main components involved in setting up
each phase. Notice that the Acquisition Phase setup is
required, but the Scoring Phase setup is optional.

Match Rule

Acquisition Phase
(Required)

Scoring Phase
(Optional)

Select Attributes

Select Attributes

Select Transformation Functions

Select Transformation Functions


Set Weightings
Set Scores
Set Thresholds

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 11

Figure 4: DQM Engine Phases with Associated Match


Rule Setups

There are seven steps to follow when creating a match rule.


The steps remain basically the same no matter which of the
three types of match rules is being created, however the
strategy behind creating the match rule may change.
Next, we will go through the seven steps of creating a match
rule. While describing the seven steps, we will create a match
rule of type Search called DUPLICATE SEARCH.
Step 1: Choose Match Rule Type
The type of match rule being created needs to be selected. The
three types of match rules that can be created are: Search,
Expanded and Bulk.
The DUPLICATE SEARCH match rule type will be Search.
Step 2: Choose Attributes
Next, determine which attributes are important when
searching for similar records. The attributes selected here will
be the basis for identifying duplicates.
For the DUPLICATE SEARCH match rule the following
attributes are selected:

Attribute
Organization
Name
URL
Phone Number
Street Address
City
Step 3: Choose Attribute Usages
The next step is to determine how the DQM engine will use
each chosen attribute. Recall that the DQM process of
identifying duplicates consists of an acquisition and scoring
phase, therefore each of the chosen attributes can be used in
the acquisition phase, scoring phase or both. As a general
rule, the use of Acquisition and Scoring will be selected for
all chosen attributes.

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 12

For the DUPLICATE SEARCH match rule the chosen


attributes will have these uses:

Attribute

Use

Organization
Name

Acquisition and
Scoring

URL

Acquisition and
Scoring

Phone Number

Acquisition and
Scoring

Street Address

Acquisition and
Scoring

City

Acquisition and
Scoring

Step 4: Assign Transformation Functions to Attributes


Remembering that transformation functions are helpful when
trying to identify duplicate records, the next step is to decide
which transformation functions should be used for each
attribute in this match rule. Each attribute/use combination
must have at least one transformation function.
There will be four different transformation functions used in
this match rule. They are:

EXACT Defined on Page 6

CLEANSE Defined on Page 6

WR(ORG)+CLEANSE Combines the CLEANSE


transformation function with the Organization word
replacement list

WR(ADD)+CLEANSE Combines the CLEANSE


transformation function with the Address word
replacement list.

For the DUPLICATE SEARCH match rule the following


transformation functions will be selected for each
Attribute/Use combination. Notice that for the attribute
Organization Name for the Scoring use, two transformation
functions have been selected:

Attribute

Use

Transformation

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 13

Organization
Name

Acquisitio
n

WR(ORG) +
CLEANSE

Scoring

WR(ORG) + EXACT
WR(ORG) +
CLEANSE

URL

Phone Number

Street Address

City

Acquisitio
n

EXACT

Scoring

EXACT

Acquisitio
n

EXACT

Scoring

EXACT

Acquisitio
n

WR(ADD) +
CLEANSE

Scoring

WR(ADD) +
CLEANSE

Acquisitio
n

CLEANSE

Scoring

CLEANSE

At this point the required acquisition phase setup is complete.


If any of the selected attributes are selected for the use of
scoring, then additional setup steps need to be completed.
Weightings and Scores need to be assigned and a Match
Threshold needs to be set.
Step 5: Assign Weightings
In addition to assigning the transformation functions,
weightings must be given to each scoring transformation
function. Weightings can further control the matching process
by giving only a percentage of the total points (Step 6: Assign
Scores) that will be assigned to the attribute. For example, a
weighting of 100% would likely be given to an EXACT
transformation function, while maybe only 75% would be given
to a CLEANSE transformation function. This means that we
want to give more points to an attribute that matches exactly,
than if it only matched after the attribute values had been
cleansed.

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 14

For the DUPLICATE SEARCH match rule the following


weightings have been assigned to each of the scoring
transformation functions:

Attribute

Use

Transformation

Organization
Name

Acquisitio
n

WR(ORG) +
CLEANSE

Scoring

WR(ORG) + EXACT

100%

WR(ORG) +
CLEANSE

75%

URL

Phone Number

Street Address

City

A Score is a number given


to an attribute in a match
rule that identifies the
importance
of
that
attribute when identifying
matching records.
The
higher the score the more
helpful that attribute is in
the duplicate identification
process.

Acquisitio
n

EXACT

Scoring

EXACT

Acquisitio
n

EXACT

Scoring

EXACT

Acquisitio
n

WR(ADD) +
CLEANSE

Scoring

WR(ADD) +
CLEANSE

Acquisitio
n

CLEANSE

Scoring

CLEANSE

Weigh
t

100%

100%

100%

100%

Step 6: Assign Scores


A Score must to be assigned to each attribute used in the
scoring phase. These scores will be used to determine how
close of a match each record in the work unit is when
compared to the input record. Scoring Attributes that provide
the most influence when determining a duplicate record should
be given the highest scores, while generic scoring attributes
are given the lowest scores.
When assigning scores, distribute 100 points across your
selected scoring attributes. Distributing scores out of 100
points is not required, however it is recommended. Based on
this logic, the following Scores have been assigned to each of
the scoring attributes for the DUPLICATE SEARCH match
rule:

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 15

Attribute

Use

Scor
e

Organization
Name

Scoring

20

URL

Scoring

30

Phone Number

Scoring

25

Street Address

Scoring

15

City

Scoring

10

TOTAL POSSIBLE SCORE


The Matching Score is
determined
by
adding
together the scores derived
from each attribute from the
work
unit
record
that
matches the same attribute
from the input value.

The Match Threshold is a


value
used
in
DQM
matching to control which
records will be brought
back as potential duplicates.
Only records with a total
score that equals or exceeds
the match threshold are
selected as a match to the
input record.

100

The purpose of assigning scores is to determine a Matching


Score for each record in the work unit during the scoring
phase. The Matching Score is determined by adding together
the scores for each attribute that matches the input record.
The Matching Score will be used to determine how close of a
match that particular record is to the input record.
Step 7: Set Thresholds
The final step is to set the Match Threshold. The value of the
Match Threshold will determine which records from the work
unit are returned to the user as potential duplicates. If a
record in the work unit falls below the Match Threshold, it is
not returned as a potential duplicate. If a record in the work
unit is equal to or greater than the Match Threshold, it is
returned as a potential duplicate.
Comparing the records in the work unit against the Match
Threshold will change depending on if DQM is being used for
Smart Search, Duplicate Prevention or Duplicate Identification.
For Smart Search and Duplicate Prevention the Threshold
Score needs to be calculated. For Duplicate Identification the
Matching Score can be compared directly to the Match
Threshold.
To calculate the Threshold Score use the equation below:
Threshold Score = (Matching Score / Total Possible Score) *
100

The Total Possible Score


is the maximum score that a
record in the work unit can
receive. This number will
vary based on the number of
attributes passed to the
DQM engine via the input
data.

To determine the Total Possible Score, add up the scores, as


defined in the match rule, for each attribute passed by the
input data. This value will change based on the number of
attributes used in the input data.

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 16

For Example:
If Organization Name and Phone Number were passed as
the input value then the Total Possible Score will be 45.
However, if Phone Number, Street Address, and City
were passed as the input value then the Total Possible Score
would be 50.
Important Note: If running a System Duplicate Identification
Batch, the Threshold Score does not need to be calculated.
The Matching Score can be compared directly to the Match
Threshold to determine if the record will be returned or not.
For the DUPLICATE SEARCH match rule the Match
Threshold will be set to 70.
It is important to remember that there is a great deal of
strategy involved when creating match rules. A deploying
company will select different attributes, scores and thresholds
depending on its business requirements, what its definition of a
matching record is and how comprehensive the match rule
needs to be.

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 17

DQM In the E-Business Suite and Customer Hub Solution


The application of the DQM search engine can be categorized
into three different areas.
1. Smart Search
2. Duplicate Prevention
3. Duplicate Identification

Smart Search

Smart Search integrates with various E-Business Suite


applications to enable a deploying company to enhance party
search screens using match rules of type Search. This
improved method of searching the TCA registry applies DQM
logic to return a list of similar records, catching records that
are misspelled and/or aliased. In addition, the search fields
exposed to the user can be configured based on the match rule
definition.
Duplicate Prevention

Duplicate Prevention enables a deploying company to


discourage duplicate records from entering the TCA registry.
Match rules of type Search are used to prevent users from
entering duplicate records through party creation screens
while match rules of type Bulk are used to prevent duplicate
records from entering the TCA registry through bulk data
import.
Figure 5 below shows that with the DQM engine applied, there
is a barrier that prevents duplicate records from entering the
TCA Registry while allowing non-duplicate records to be
created.
DUPLICATE
PREVENTION
Bulk
Import
Duplicates
Party
Creation
TCA REGISTRY

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 18

Figure 5: DQM Prevents Duplicate Records from


Entering the TCA Registry

Duplicate Identification

Duplicate Identification allows the deploying company to


identify existing duplicate records within the TCA registry
using either match rules of type Extended or Bulk.
Duplicate Identification searches are run in batch mode and
the results are returned in duplicate sets that can be evaluated
and resolved as necessary.

As seen from the three aforementioned areas of application,


the DQM engine provides the power to combat duplicate
records in multiple ways. Figure 6 below shows a
comprehensive graphic of all applications of the DQM search
engine, demonstrating the complete DQM solution for
SMART
SEARCH
DUPLICATE
PREVENTION
Bulk
Import
Duplicates

DUPLICATE
IDENTIFICATI
ON

Party
Creation
TCA REGISTRY

managing duplicate records within the TCA customer registry.


Figure 6: Three Applications of the DQM Engine

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 19

Conclusion
The issue of information quality has become a significant
obstacle in establishing a single view of customer information
and gaining accurate, reliable data for decision making. The
Oracle Data Quality Management engine will help arm a
deploying company with the tools necessary to achieve a single
view of its customers by enabling the identification and
prevention of duplicate records in the TCA registry. Realizing
that no two companies face exactly the same data quality
issues, the flexibility of the DQM engine allows a custom
solution for each deploying company.
DQM is one piece of TCAs comprehensive data quality solution
which was created to facilitate the sustained commitment to
managing your data quality over time and adjusting processes
to the ongoing changes in your business environment. Within
the Oracle E-Business Suite, you can build meaningful and
powerful rules to search for registry information, find matches,
identify duplicates, merge them and proactively prevent the
entry of duplicates at the front end. By providing you with
flexible and powerful Data Quality Management tools, you as
the deploying company will have a better answer to the
question Who is my customer? This not only facilitates
important strategic decisions, but empowers you to more
confidently meet the challenges of safeguarding the integrity of
your trading community data one of the most vital resources
necessary for succeeding in todays competitive global
economy.
Within the deploying organization there is typically a single
person or a team of people responsible for managing data
quality. Oracle refers to this role as the Data Librarian. To
facilitate this job role the Customer Data Librarian (CDL)
application has been created, which combines the duplicate
identification (DQM) and duplicate resolution (Party Merge)
tools into a user-friendly html interface created specifically for
individuals and teams responsible for an organizations
information quality.

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 20

Appendix A
DQM Examples
This section will give an example of the application of the DQM
engine in the following three areas:
1) Smart Search
Where the DQM engine is leveraged by E-Business Suite
applications to enhance TCA registry search screens
2) Duplicate Prevention
Where the DQM engine is leveraged by E-Business Suite
applications to prevent duplicates from entering the
TCA registry through party creation screens
3) Duplicate Identification
Where the DQM engine is used to identify existing
duplicate records in the TCA registry using System
Duplicate Identification Batches

Smart Search Example

In the following example the DUPLICATE SEARCH match


rule will be used. Refer back to the section titled: Defining
Match Rules (Page 11) for a definition of this match rule. This
example represents a real situation where the DQM engine is
used by an E-Business Suite application to enhance a party
search screen. In this scenario a user tries to search for an
existing party called Business World.
Assume the following two records already exist in the TCA
registry. Record #1 is misspelled, while record #2 is simply
another Business World in a different city.
EXISTING TCA DATA
#1
#2

Org Name

URL

Street

City

Bisness

www.bw.co

Address
1977
Main

San

World
Business

m
www.bw.co

Street
22
Roosevelt

Jose
Detroit

World

Blvd

Input Data
A user tries to search for an organization party and enters the
information below:
INPUT DATA
Org Name

URL

Street

City

Business

www.bw.co

Address
1977
Main

San

World

Street

Jose

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 21

Transformed Input Data


The DQM engine must first transform the input data provided
by the user. Each attribute of the input data is transformed
using the superset of transformation functions defined in the
match rule. Recall that the transformation functions for the
match rule DUPLICATE SEARCH are as follows:
Attribute

Use

Transformation

Organization

Acquisitio

WR(ORG) +

Name

n
Scoring

CLEANSE + EXACT
WR(ORG)

URL

Acquisitio

EXACT

n
Scoring

EXACT

Acquisitio

EXACT

n
Scoring

EXACT

Acquisitio

WR(ADD) +

n
Scoring

CLEANSE +
WR(ADD)

Acquisitio

CLEANSE
CLEANSE

n
Scoring

CLEANSE

Phone Number
Street Address
City

Using the superset of these transformation functions, the


transformed input data looks like this:
TRANSFORMED INPUT DATA
Org Name

URL

Street Address

City

WR(ORG)+EXACT

EXACT

WR(ORG)+CLEANSE

CLEANSE

BUSINESS

WWW.BW.COM

1977 MN ST

SN JS

WORLD
WR(ORG)+CLEANSE

BSNS WRLD

In the Acquisition Phase


the DQM engine acquires a
subset of matching records
based on the acquisition
attributes selected in the
match rule. This subset is
called the Work Unit.

Acquisition Phase
In the Acquisition Phase, the DQM engine compares the
transformed input data to the transformed TCA data (Staging
Tables) to find matching records. The DQM engine only uses
those attributes and transformation functions that are defined
in the selected match rule for use in the Acquisition Phase.
Based on the definition of the match rule DUPLICATE
SEARCH, the table below shows the transformed input data
that will be used for the Acquisition Phase. It also shows the
transformed TCA data, records #1 and #2, with which the
input data will be matched against.
Matching values are outlined with black boxes.
TRANSFORMED INPUT DATA

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 22

Org Name

URL

Street Address

City

WR(ORG)+CLEANSE

EXACT

WR(ORG)+CLEANSE

CLEANSE

BSNS WRLD

WWW.BW.COM

1977 MN ST

SN JS

TRANSFORMED TCA DATA (Staging Table)

#1

Org Name

URL

Street Address

City

WR(ORG)+EXACT

EXACT

WR(ORG)+CLEANSE

CLEANSE

BISNESS WORLD

WWW.BW.COM

1977 MN ST

SN JS

WR(ORG)+EXACT

EXACT

WR(ORG)+CLEANSE

CLEANSE

BUSINESS

WWW.BW.COM

22 RSVLT BLVD

DTRT

WR(ORG)+CLEANSE

BSNS WRLD

#2

WORLD
WR(ORG)+CLEANSE

BSNS
SinceWRLD
both records #1 and #2 had attributes that matched the

input data, they are returned to the Work Unit for evaluation
during the Scoring Phase.

Scoring Phase
In the scoring phase, each record identified for the Work Unit
by the Acquisition Phase is evaluated and receives a Matching
Score. To calculate the Matching Score the DQM engine only
uses the attributes and transformation functions that are
defined in the selected match rule for use in the Scoring Phase.
With that in mind, the table below shows the records in the
Work Unit being evaluated against the input data with the end
result being a Matching Score.
TRANSFORMED INPUT DATA
Org Name

URL

Street Address

City

WR(ORG)+EXACT

EXACT

WR(ORG)+CLEANSE

CLEANSE

BUSINESS

WWW.BW.COM

1977 MN ST

SN JS

WORLD
WR(ORG)+CLEANSE

BSNS WRLD
Work Unit
Org Name (20)

URL (30)

Street Address

City (10)

(15)
#1

WR(ORG)

EXACT(100%)

WR(ORG)

CLEANSE(10

+EXACT(100%)

WWW.BW.COM

+CLEANSE(100%)

0%)

1977 MN ST

SN JS

BISNESS WORLD

Matchin

WR(ORG)
+CLEANSE(75%)

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 23

Score

70

#2

WR(ORG)

EXACT(100%)

WR(ORG)

CLEANSE(10

+EXACT(100%)

WWW.BW.COM

+CLEANSE(100%)

0%)

22 RSVLT BLVD

DTRT

BUSINESS
WORLD

Recall that the Matching Score is determined by adding


together the scores for each attribute that matches the input
data on one of the transformation functions defined in the
match rule. Below shows how the Matching Scores in the table
above were determined.
The Matching Score for record #1 is:
Org Name (20 x .75)* + URL (30) + Street Address (15) +
City (10) = 70
The Matching Score for record #2 is:
Org Name (20) + URL (30) = 50
* The score given to Org Name for the first record is only 75% because
the attribute matched on the WR(ORG)+CLEANSED transformation
function. This was determined when the weightings were set in the match
rule.

Apply Match Threshold


Next, in order to determine if these records will be returned to
the user as potential duplicates, the Threshold Score needs to
be calculated and compared to the threshold set in the Match
Rule. Recall that to calculate the Threshold Score, the
#1
following equations is used:
Threshold Score = (Matching Score / Total Possible Score) x
100
The Matching Score has already been calculated in the Scoring
Phase so the only value left to be determined is the Total
Possible Score. To determine the Total Possible Score, add up
the scores, as defined in the match rule, for each attribute that
is passed a value by the input data. Recall that the scores were
defined as the following:
Attribute

Use

Scor
e

Organization

Scorin

Name

20

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 24

50

URL

Scorin

30

g
Phone Number

Scorin

25

g
Street Address

Scorin

15

g
City

Scorin

10

Since the user input Org Name, URL, Street Address and City
the Total Possible Score is the sum of the scores defined for
each of these attributes. In this example the Total Possible
Score is:
Org Name (20) + URL (30) + Street Address (15) + City
(10) = 75
Now that the Matching Score and the Total Possible Score are
known, the Threshold Score can be calculated. It is as follows
for each of the records:
The Threshold Score for record #1 is:
(70 / 75) * 100 = 93.3
The Threshold Score for record #2 is:
(50 / 75) * 100 = 66.7
Finally, the DQM engine will only return those records that
meet or exceed the Match Threshold, which as you recall, is
70. So in this example only record #1 meets or exceeds the
Match Threshold.

Results
The user tried to search for the following record:
USER ENTERED DATA
Org Name

URL

Street

City

Business

www.bw.co

1977 Main

San

World

Street

Jose

Using the DUPLICATE SEARCH match rule, the DQM engine


returned only record #1, to the user:
MATCHING RECORDS
Org Name

URL

Street
Address

City

THRESHOL
D SCORE

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 25

#1

Bisness

www.bw.co

1977 Main

San

World

Street

Jose

93.3

Next Steps
At this point the user can:

Continue their process with the DQM engines


suggested party

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 26

Duplicate Prevention Example

In the following example the DUPLICATE SEARCH match


rule will again be used. Refer back to the section titled:
Defining Match Rules (Page 11) for a definition of this match
rule. This example represents a real situation where the DQM
engine is used by an E-Business Suite application to prevent a
duplicate record from entering the TCA registry through a
party creation screen.
In this scenario a user tries to create a new party called Vision
Corporation through a party create screen, but before the
party is created, the DQM engine uses the DUPLICATE
SEARCH match rule to search the TCA registry for matching
records and return a list of potential duplicates to the user.
Assume the following four records already exist in the TCA
registry.
EXISTING TCA DATA
#1
#2
#3
#4

Org Name

Phone

Street Address

City

Vision Corp

503.554.256

12344 Hammond

Portlan

Corp Corp.
Vission

8
503-553-

Blvd. Shine Road


1552

d
Portlan

Viasion

2669 554(503)

1234 Hammond

d
Portlan

Corporation
Visin

2568
503.332.326

Boulevard
1956
Brookside Dr.

d
Portlan

Corporation

Input Data
A user tries to create an organization party and enters the
information below. The user accidentally types Visin instead
of the properly spelled Vision:
INPUT DATA
Org Name

Phone

Street Address

City

Visin

503-554-

1234 Hammond

Portlan

Corporation

2568

Blvd.

Transformed Input Data


The DQM engine must first transform the input data provided
by the user. Each attribute of the input data is transformed
using the superset of transformation functions defined in the
match rule. Recall that the transformation functions for the
match rule DUPLICATE SEARCH are as follows:
Attribute

Use

Transformation

Organization

Acquisitio

WR(ORG) +

Name

n
Scoring

CLEANSE
WR(ORG) + EXACT

URL

Acquisitio

EXACT

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 27

Phone Number
Street Address
City

Scoring

EXACT

Acquisitio

EXACT

n
Scoring

EXACT

Acquisitio

WR(ADD) +

n
Scoring

CLEANSE +
WR(ADD)

Acquisitio

CLEANSE
CLEANSE

n
Scoring

CLEANSE

Using the superset of these transformation functions, the


transformed input data looks like this:
INPUT DATA
Org Name

Phone

Street Address

City

WR(ORG)+EXACT

EXACT

WR(ORG)+CLEANSE

CLEANSE

VISIN CORP

5035542

1234 HMND BLVD

PRTLND

568

WR(ORG)+CLEANSE

VSN CORP

Acquisition Phase
In the Acquisition Phase, the DQM engine compares the
transformed input data to the transformed TCA data (Staging
Tables) to find matching records. The DQM engine only uses
those attributes and transformation functions that are defined
in the selected match rule for use in the Acquisition Phase.
Based on the definition of the match rule DUPLICATE
SEARCH, the table below shows the transformed input data
that will be used for the Acquisition Phase. It also shows the
transformed TCA data, records #1, #2, #3, and #4, with which
the input data will be matched against.
Matching values are outlined with black boxes.
TRANSFORMED INPUT DATA
Org Name

Phone

Street Address

City

WR(ORG)+CLEANSE

EXACT

WR(ORG)+CLEANSE

CLEANSE

VSN CORP

503554256

1234 HMND BLVD

PRTLND

8
TRANSFORMED TCA DATA (Staging Table)

#1

Org Name

Phone

Street Address

City

WR(ORG)+EXACT

EXACT

WR(ORG)+CLEANSE

CLEANSE

VISION CORP

503554256

12344 HMND BLVD

PRTLND

WR(ORG)+CLEANSE

VSN CORP

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 28

#2

WR(ORG)+EXACT

EXACT

WR(ORG)+CLEANSE

CLEANSE

VISSION CORP

503553266

1552 SHN RD

PRTLND

WR(ORG)+CLEANSE

VSN CORP

#3

WR(ORG)+EXACT

EXACT

WR(ORG)+CLEANSE

CLEANSE

VIASION CORP

503554256

1234 HMND BLVD

PRTLND

WR(ORG)+CLEANSE

VSN CORP

#4

WR(ORG)+EXACT

EXACT

WR(ORG)+CLEANSE

CLEANSE

VISIN CORP

503332326

1956 BRKSD DR

PRTLND

WR(ORG)+CLEANSE

VSN CORP

Since records #1, #2, #3, and #4 had attributes that matched
the input data, they are returned to the Work Unit for
evaluation during the Scoring Phase.

Scoring Phase
In the scoring phase, each record identified for the Work Unit
by the Acquisition Phase is evaluated and receives a Matching
Score. To calculate the Matching Score the DQM engine only
uses the attributes and transformation functions that are
defined in the selected match rule for use in the Scoring Phase.
With that in mind, the table below shows the records in the
Work Unit being evaluated against the input data with the end
result being a Matching Score.
TRANSFORMED INPUT DATA
Org Name

Phone

Street Address

City

WR(ORG)+EXACT

EXACT

WR(ORG)+CLEANSE

CLEANSE

VISIN CORP

503554256

1234 HMND BLVD

PRTLND

WR(ORG)+CLEANSE

VSN CORP
Work Unit
Org Name (20)

#1

Phone

Street Address

City (10)

Matchin
Score

(25)

(15)

WR(ORG)

EXACT(100%)

WR(ORG)

CLEANSE(10

+EXACT(100%)

503554256

+CLEANSE(100%)

0%)

VISION CORP

12344 HMND BLVD

PRTLND

WR(ORG)
+CLEANSE(75%)

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 29

50

#2

WR(ORG)

EXACT(100%)

WR(ORG)

CLEANSE(10

+EXACT(100%)

503553266

+CLEANSE(100%)

0%)

VISSION CORP

1552 SHN RD

PRTLND

+CLEANSE(75%)
WR(ORG)

EXACT(100%)

WR(ORG)

CLEANSE(10

+EXACT(100%)

503554256

+CLEANSE(100%)

0%)

VIASION CORP

1234 HMND BLVD

PRTLND

+CLEANSE(75%)
WR(ORG)

EXACT(100%)

WR(ORG)

CLEANSE(10

+EXACT(100%)

503332326

+CLEANSE(100%)

0%)

VISIN CORP

1956 BRKSD DR

PRTLND

25

WR(ORG)

#3

65

WR(ORG)

#4

WR(ORG)
+CLEANSE(75%)

Recall that the Matching Score is determined by adding


together the scores for each attribute that matches the input
data on one of the transformation functions defined in the
match rule. Below shows how the Matching Scores in the table
above were determined.
The Matching Score for record #1 is:
Org Name (20 x .75)* + Phone (25) + City (10) = 50
The Matching Score for record #2 is:
Org Name (20 x .75)* + City (10) = 25
The Matching Score for record #3 is:
Org Name (20 x .75)* + Phone (25) + Street Address (15) +
City (10) = 65
The Matching Score for record #4 is:
Org Name (20) + City (10) = 30
* The score given to Org Name for the records #1, #2, and #3 is only
75% because the attribute matched on the WR(ORG)+CLEANSED
transformation function. This was determined when the weightings were
set in the match rule.

Apply Match Threshold


Next, in order to determine if these records will be returned to
the user as potential duplicates, the Threshold Score needs to
be calculated and compared to the threshold set in the Match
Rule. Recall that to calculate the Threshold Score, the
following equation is used:
Threshold Score = (Matching Score / Total Possible Score) x
100

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 30

30

The Matching Score has already been calculated in the Scoring


Phase so the only value left to be determined is the Total
Possible Score. To determine the Total Possible Score, add up
the scores, as defined in the match rule, for each attribute that
is passed a value by the input data. Recall that the scores were
defined as the following:

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 31

Attribute

Use

Scor
e

Organization

Scorin

Name

URL

Scorin

20

30

g
Phone Number

Scorin

25

g
Street Address

Scorin

15

g
City

Scorin

10

Since the user input Org Name, Phone, Street Address and City
the Total Possible Score is the sum of the scores defined for
each of these attributes. In this example the Total Possible
Score is:
Org Name (20) + Phone (25) + Street Address (15) + City
(10) = 70
Now that the Matching Score and the Total Possible Score are
known, the Threshold Score can be calculated. It is as follows
for each of the records:
The Threshold Score for record #1 is:
(50 / 70) * 100 = 71.4
The Threshold Score for record #2 is:
(25 / 70) * 100 = 35.7
The Threshold Score for record #3 is:
(65 / 70) * 100 = 92.9
The Threshold Score for record #4 is:
(30 / 70) * 100 = 42.9
Finally, the DQM engine will only return those records that
meet or exceed the Match Threshold, which as you recall, is
70. So in this example records #1 and #3 meet or exceed the
Match Threshold.

Results
The user tried to create the following record:
USER ENTERED DATA

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 32

Org Name

Phone

Street Address

City

Visin

503-554-

1234 Hammond

Portlan

Corporation
2568
Blvd. match rule, thedDQM engine
Using the DUPLICATE
SEARCH

returned the following list of records to the user as potential


duplicates:

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 33

POTENTIAL DUPLICATES
Org Name

Phone #

Street Address

City

THRESHOL
D SCORE

#3

Viasion

(503) 554-

1234 Hammond

Portlan

#1

Corporation

2568

Boulevard

Vision Corp

503.554.256

12344 Hammond

Portlan

Blvd.

Next Steps
At this point the user can exercise one of two options:

Continue creating the new party

Select one of the DQM engines suggested parties and


update the existing record

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 34

92.9

71.4

Duplicate Identification Example

Duplicate Identification is the process of searching the TCA


registry for existing duplicate records. A System Duplicate
Identification Batch(SDIB) is submitted, and returned are sets
of duplicate records. Since match rules of type Search
cannot be used for Duplicate Identification, the DUPLICATE
SEARCH match rule created previously in this document
cannot be used. Instead, a match rule of type Expanded with
a name of DUPLICATE IDENTIFICATION will be used. For
simplicity, the details of this match rule will not be exposed.
However, assume that it functions similar to the DUPLICATE
SEARCH match rule.
System Duplicate Identification Batch
To execute the Duplicate Identification process, an SDIB needs
to be submitted to search the TCA registry for existing
duplicate parties. To create the SDIB, select either a match
rule of type Expanded or Bulk and submit. In this case the
DUPLICATE IDENTIFICATION expanded type match rule will
be used.
The Input Record is the
record that the DQM engine
is trying to find duplicates
for. In Smart Search and
Duplicate Prevention the
Input Record comes from a
user, however in Duplicate
Identification
it
is
determined randomly from
the TCA registry or subset
of the TCA registry.

Input Record
Unlike the Smart Search and Duplicate Prevention methods,
Duplicate Identification does not have a user input. Therefore,
for the input record the DQM engine randomly selects a party
from the TCA registry and evaluates the rest of the registry for
duplicates. During the SDIB creation process the user has the
option to limit the duplicate identification process to a subset
of the TCA registry. This matching process continues until all
parties in the TCA registry, or subset, have been evaluated.

Acquisition and Scoring Phases


The acquisition and scoring phases work the same as in the
Smart Search and Duplicate Prevention methods with one
exception. When comparing records in the work unit to the
Match Threshold, the DQM engine uses the Matching Score
instead of calculation the Threshold Score. Refer back to Step
7 in the section Defining Match Rules for a description of
this process.

Results
Below is a sample of a few duplicate sets that DQM would
return after running an SDIB with the expanded match rule
DUPLICATE IDENTIFICATION:
MASTER RECORD

POSSIBLE DUPLICATES

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 35

Gorman Manufacturing

gormmun mfg.
Gorman
Manufacturing

MASTER RECORD
Business World

POSSIBLE DUPLICATES
Bisniss Wurld
Business world
bussines worlld

Next Steps
At this point the user can exercise one of two options:

Determine that the parties are not duplicates and do


nothing

Create a merge request to have the duplicate parties


merged

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 36

What To Do With Duplicates


After duplicate records have been identified by the DQM
engine, what next? Listed below are a few things that a user
can choose to do:

1) Create a Relationship
The user may find that the potential duplicates are not
exact duplicates, but that they are related records. In
this case, a relationship can be created to reflect the
connection the two parties have.
Party Merge is used to
consolidate
duplicate
records into one master
record.
All transactional
data associated with the
duplicate party is re-pointed
to the master party.
Account Merge is used to
consolidate accounts from
either the same party or
different parties. Like party
merge,
all
transactional
data associated with the
merged account is repointed to the master
account.

2) Merge Parties
If records are found to be duplicate, the parties can be
merged using the Party Merge functionality.

3) Merge Accounts
If two duplicate accounts are identified they can be
merged using the Account Merge functionality.

4) Create a Merge Request


If DQM finds duplicate records but the user does not
have the ability to merge, he/she can submit a Merge
Request which will notify the Data Librarian that the
records should be merged.

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 37

Appendix B
Definitions and Terminology
In order to minimize confusion that can arise due to differences
in interpretation of terminology, the following definitions are
provided for terms that are used throughout this document.

Account refers to the details of the deploying companys


selling relationship with a particular Party.

Account Merge refers to the process of resolving


duplicate customer account records. Account Merge is used
to merge transactions from a source customer account to a
target customer account. Customer Account Merge only
merges the customer accounts and related transactions. It
does not merge any TCA registry level information. You
must separately run Party Merge in order to merge the TCA
registry level information about the party if it is
appropriate.

Acquisition Attributes are used for selecting the initial,


most relevant subset of records for matching, forming what
is called the Work Unit.

Attribute refers to a column in a TCA registry.


Attributes are used by DQM to construct match rules and
identify potential duplicates.

Customer is used to represent a party with whom the


deploying company has a selling relationship, regardless of
whether anything has actually been purchased or serviced.

Data Librarian refers to the employee or consultant for


the deploying company who is responsible for identification
and resolution of duplicate or incomplete party or customer
account information in their implementation of the Oracle
E-Business Suite. This role could be filled by a single
person or a team of people.

Deploying Company refers to the Oracle customer that


has, or will, install, implement, and run all or part of The
Oracle E-Business Suite or Customer Data Hub solution.

Duplicate is a record that a user has deemed a


duplicate. In this case the duplicate record should be
merged using Party Merge and if necessary Account Merge.

Implementer refers to either an internal or external


employee working for the deploying company to implement

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 38

any or all of the E-Business Suite or Customer Data Hub


solution.

Match Rules are used to determine if two parties are the


same and should be considered as a match. For example,
you might want to setup a match rule that says: If the
party name, state and primary phone number exactly
match, then the records should be returned as a match.

Merge refers to an operation in which one entity (the


merge-from entity) is mapped to another entity of the same
type (the merge-to entity), then any child entities of the
merge-from entity are transferred, copied, or merged to the
merge-to entity. The merge-from entity is inactivated or
deleted at the option of the user. Example: merging one
party with another.

Party is an entity in the Trading Community Model that


can enter into business relationships. A party is a person,
organization (branch, subsidiary, legal entity, holding
company, etc) or relationship (this is a relationship between
two parties).

Party Merge refers to the process of resolving duplicate


TCA registry records. With the Party Merge feature you
can consolidate duplicate parties, integrate an acquired
party into another party, or consolidate duplicate party sites
or contacts of a party in the TCA party registry. The related
child entities that get merged or transferred include party
relationships, contact information, party profiles, customer
accounts, and information obtained from third-party
sources.

Potential Duplicate is a record that might be a


duplicate, but the user needs to make the final
determination if it is in fact a duplicate.

Scoring Attributes are used to calculate a score for


each record in the Work Unit.

System Duplicate Identification Batch is a process


that can be run by the deploying company based upon the
match rules that they have defined as relevant to their
organization. This process evaluates the records in the TCA
registry to determine which are potential duplicates of each
other. Commonly used by Data Librarians to cleanse the
customer database by identifying existing duplicate
records.

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 39

Thresholds are cutoffs that are used to exclude records


that dont have high enough scores to be considered as
potential duplicates. For example, you might want to set up
a threshold that states: if the total match score is less than
50, do not show the record as a potential duplicate.

Transformation Function transforms or changes


attributes into different representations for the purposes of
identifying potential duplicate records. They are helpful in
instances where words are aliased or misspelled.
Transformation Functions can call upon Word Replacement
Lists.

Word Replacement Lists are word mappings that are


treated as equivalents for matching. They help to
standardize data where nicknames or abbreviations may
have been used. They are optionally used by
Transformation Functions. For example, you might want to
map Robert, Rob, Robbie, Bob and Bobbie as all equivalent
to one another for the purposes of matching.

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 40

Appendix C
An advantage to The Oracle solution for data quality is the
ability to not only identify duplicate parties, but also merge
them based an intimate knowledge of the Trading Community
Architecture data model. This knowledge allows the solution to
accurately combine all associated transactions and
relationships tied to any of the duplicates based on registered
routines created directly by the experts in each of the EBusiness application teams. The following is a dated list of the
Oracle E-Business Suite Applications that have registered
special routines to ensure that all associated data is merged
properly and data integrity is maintained at the highest level.
E-Business Suite Applications That Have Registered Merge Routines
with Party Merge
as of 9/2003

ACTIVITY BASED MANAGEMENT


ADVANCED PRICING
ADVANCED SERVICE ONLINE
APPLICATIONS OBJECT LIBRARY
COLLECTIONS
COMPLEX MAINTENANCE, REPAIR AND OVERHAUL
CONTRACT FOR RIGHTS
CONTRACTS CORE
CRM FOUNDATION
CRM FOUNDATION INTERACTION HISTORY
CRM FOUNDATION RESOURCE MANAGER
CRM FOUNDATION TASKS AND NOTES
CRM FOUNDATION TERRITORIES
CRM FOUNDATION USER MANAGEMENT
CUSTOMER CARE
GRANTS PROPOSAL
HUMAN RESOURCES
INSTALLED BASE
ISTORE
LEASE MANAGEMENT
MARKETING
MOBILE
OFFERS
ORDER CAPTURE
ORDER MANAGEMENT
PARTNERS ONLINE
PARTNER RELATIONSHIP MANAGEMENT (PREVIOUSLY PARTNERS ONLINE)
PROJECTS

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 41

PROJECT CONTRACTS
PROPOSALS
RECEIVABLES
SALES
SERVICE
SERVICE CONTRACTS
SHIPPING
SOURCING
SPARES MANAGEMENT
TCA (TRADING COMMUNITY ARCHITECHTURE)
XML GATEWAY

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 42

Appendix D
An advantage to The Oracle solution for data quality is the
ability to merge customer accounts based an intimate
knowledge of the Trading Community Architecture data model.
This knowledge allows the solution to accurately combine all
associated transactions and relationships tied to any of the
duplicate accounts based on registered routines created
directly by the experts in each of the E-Business application
teams. The following is a dated list of the Oracle E-Business
Suite Applications that have registered special routines to
ensure that all associated data is merged properly and data
integrity is maintained at the highest level.
E-Business Suite Applications Referenced by Customer Account Merge
as of 9/2003

ADVANCED PRICING
COLLECTIONS
CONTRACTS CORE
CRM FOUNDATION INTERACTION HISTORY
CRM FOUNDATION TASK
CUSTOMER CARE
FEDERAL FINANCIALS
FINANCIALS LATIN AMERICA LOCALIZATION
GLOBAL ACCOUNTING ENGINE
GLOBAL FINANCIAL FINANCIALS COMMON COUNTRY
GRANTS ACCOUNTING
INSTALLED BASE
INVENTORY
ISTORE
MARKETING
MASTER SCHEDULE AND PLANNING
MULTI-CURRENCY CREDIT CHECKING
ORDER CAPTURE
ORDER MANAGEMENT
PROJECTS
PROPERTY MANAGER
PUBLIC SECTOR
PURCHASING
RELEASE MANAGEMENT
SALES COMPENSATION
PAYABLES
PROVISIONING
RECEIVABLES
SERVICE

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 43

SERVICE CONTRACTS
SHIPPING EXECUTION
SPARES MANAGEMENT
TCA (TRADING COMMUNITY ARCHITECTURE)
TRAINING ADMINISTRATION

Oracle Trading Community Architecture Data Quality Management Understanding the Concepts Page 44

Oracle Trading Community Architecture Best Practices


Data Quality Management - Understanding the Concepts
April 2004
Contributing Authors: Matt Stitz, Alan Adams
Oracle Corporation
World Headquarters
500 Oracle Parkway
Redwood Shores, CA 94065
U.S.A.
Worldwide Inquiries:
Phone: +1.650.506.7000
Fax: +1.650.506.7200
www.oracle.com
This document is provided for information purposes only and the contents hereof are subject to change without notice. This document is not warranted to be error-free, nor
subject to any other warranties or conditions, whether expressed orally or implied in law, including implied warranties and conditions of merchantability or fitness for a
particular purpose. We specifically disclaim any liability with respect to this document and no contractual obligations are formed either directly or indirectly by this document.
This document may not be reproduced or transmitted in any form or by any means, electronic or mechanical, for any purpose, without our prior written permission.
Oracle is a registered trademark of Oracle Corporation. Various
product and service names referenced herein may be trademarks
of Oracle Corporation. All other product and service names
mentioned may be trademarks of their respective owners.
Copyright 2004 Oracle Corporation
All rights reserved.

You might also like