Sie sind auf Seite 1von 7

TCA and DQM Overview

Please click to proceed through presentation http://oracle.anipassi.com


Anil Passi
Stored independently of TCA. Users
ensure Vendor Number/name is Unlike HRMS, OSS Student System
same as in TCA simply references Student Number
these fields from TCA **Same as Party Number
Table HZ_PARTIES Last Name
Payables **Referenced from TCA
PO_VENDORS First Name
Vendor Number **Referenced from TCA
Vendor Name Date Of Birth
Employee Id **Referenced from TCA
Party Id

References

HRMS TCA
PER_ALL_PEOPLE_F HZ_PARTIES
Kept in
EMPLOYEE_NUMBER[CID] Party Number
Synch-In
LAST_NAME Last Name
Different
FIRST_NAME First Name
physical
DATE_OF_BIRTH Date Of Birth
schemas
PERSON_ID Gender
Reference to Party Id Reference to Person Id
PARTY_ID DQM Staging Process

Staging
Transformation
rules applied

Applies Filter criteria


DQM Screen Applies weightings DQM Registry
Displays Highest score records on top Applies Scoring HZ_STAGED_PARTIES

http://oracle.anipassi.com
Purpose of Transformation in DQM?

Staging ->Exact and Cleanse are common


transformations given by Oracle [Pre-Seeded]
First Name Last Name
->You can define as many custom
Anil Passian transformation rules, like replacing Z with S

Exact Cleanse Exact Cleanse ->Multiple transformation rules can be applied to


each column. For each transformation applied, a
ANIL ANL PASSIAN PSN independent column value is stored in staging
table

EXACT TRANSFORMATION
Make all characters UPPER case
Remove non-alphanumeric characters
Match Found Replace multiple white spaces by single white space

CLEANSE All above rules of EXACT plus


Remove double letters
Remove all vowels, except leading vowels

Transformation can be applied again to values that user


is searching. Lets say user searches on Aneeal Passion

ANEEAL ANL PASSION PSN ->These transformations can differ from those used for
staging. But most likely we use same transformations
when searching as that for staging.
Exact Cleanse Exact Cleanse
While doing DQM config, we will give Higher Scoring to
Aneeal Last Name, and lower Scoring for First Name match
Passion
First Name Last Name For Last Name, we give 100% Weighting to EXACT
Match, and say 65% weight to CLEANSE. Hence match
http://oracle.anipassi.com
based on Exact will have higher rank
User searches for Aneeal Passion.eeeks
To setup DQM for Staging:-
1. Decide the columns that you would like to search & score/rank upon. For example Last
Name, First Name, DOB, NI Number etc
2. Define custom attributes for those columns that are not held within TCA table like
HZ_PARTIES. For example NI Number is not in TCA tables, hence we can define Custom
Attributes for NI Number, and then define Custom Procedure that returns NI number from
HR tables [for a given party_id]
3. Define the transformations for each column. Multiple transformations can be defined

Make a
transformation
active to allow it
being used for
scoring or data
search

If Acquisition Flag is checked, it


means this transformed value will
be used for searching initial set of
Multiple transformations can be matching records.
applied to each column.

If Acquisition Flag is NOT


checked, it means this
transformed value will be used
solely for the purpose of Scoring
the records returned by
acquisition phase.
http://oracle.anipassi.com
Define match rules

Scoring happens after Filtration and


Acquisition has been done. The resultant
set of records are then run through scoring
process, so that we can display highest
ranked matches on the top of list

These are the transformations which


will be used for performing search.
Only those Transformations that have
their Acquisition Checkbox Flag
checked [in previous slide], can be
used in acquisition phase

A custom attribute can be defined to Filter


out invalid Party Records. Filter is
applicable if we can programmatically
suppress duplicate/invalid Party records
from appearing.
http://oracle.anipassi.com
Scoring Phase [After Acquisition phase is completed]

Assign score to each column.


Match found on Party Number
will have Score 100, whereas
match on Last Name can be
score of 90. Score for match
on Title[Mr/Ms] will be still
lower

Within each column, each


transformation can be given
different Weight. EXACT match
can be given 100% weight,
whereas CLEANSE [after
removing vowels] will be give
lower weight.

Transformations used in Acquisition can


be used for Scoring too.

http://oracle.anipassi.com
Records with Highest score are displayed
on the Top of DQM Screen

End of presentation - Thanks


http://oracle.anipassi.com

Das könnte Ihnen auch gefallen