Data Quality Management(DQM) in TCA

In this article I will touch base upon the data quality management feature of TCA. This article also contains a link to a DQM presentation that will compliment the Q & A in this article. A combination of this article and thepowerpoint presentation should help you understand the basics of DQM [ Data Quality Management ] in Oracle TCA. 
Is DQM a part of TCA?
Yes it is, the setup can be done using responsibility named "Trading Community Manager".

What are the setup steps for DQM?
Note: These steps will get clearer once you finish reading this article + PowerPoint Presentation
Here are the steps :-
1. Identify the Attributes that you wish to use for Search and Scoring.
2. Define custom attributes for those columns on which you would like to either search or score, but those columns are not in TCA tables.
3. Assign transformation to each column. Define custom transformation rules if need be, by using PL/SQL 
4. Decide which transformed values are used for Searching and which transformed values are to be used for [both] Searching and/or just Scoring
The attribute transformations that you wish to search upon, will have "Acquisition" checkbox set to Yes.
5. Define your match rules. You can define as many match rules
6. In Match rules, you specify how the users search criteria[values being used by users] are transformed prior to acquisition.
7. Define scoring/ranking rules
8. Run the staging process.


Give me some examples where DQM is used?
Lets say your organization has 50000 employees and a person named john smith gets recruited for a job. HRMS team must know whether this John Smith has had any kind of interaction with the organization in past. Ideally if John Smith has had any kind of interaction with the company in past, then there must exist an entry for this record in TCA. In this case, HR must ensure that HRMS record can be made to point to existing TCA record. By doing so a 360 degree view of the employee be maintained in your organization.


So DQM is about being able to identify a duplicate entry in your system.
Correct, and this I made possible by applying some transformations as you will see in the presentation.


Is DQM only applicable at party level in TCA?
No way, DQM can be made to work at party address level & contact level too.


Isn't finding duplicates as simple as matching the social security number or national insurance number?
Correct, but say if your payables department wants to pay expenses to a third party for their interview travel expenses, then payables team will not have access to SSN or "NI Number". In this case, they rely upon other attributes of the person to find duplicates.


Ah, it means DQM will not be used merely by HRMS department or the Receivables department?
Certainly not, in an efficient organization, any department that registers a party into oracle apps must have access to DQM functionality.


But what if Payables team wants to use DQM, but with different set of search and ranking criteria altogether?
This is a very common requirement. Hence in this case Oracle lets you define multiple "Match Rules" in DQM.


Fine, we can define different Match Rules, but how will we enforce Payables users to use say "Match Rule 1" and HRMS users to use say "Match Rule 2"?
You can differentiate such requirements by creating different DQM search responsibilities. For each such responsibility you will assign profile option "DQM Match Rule for Search" to respective value.
In case you do not wish to have different responsibilities, though recommended, you can set this profile option at user level too.


But DQM searches for duplication in TCA tables, however "NI Number" and SSN [Social Security] are stored in HRMS tables. How will DQM search data in other modules?
You can define custom attributes in DQM. Custom attribute is nothing but a pseudo column which uses pl/sql to return the value,in this case from per_people_f


Does DQM search data within TCA tables?
Well, after defining rules for DQM setup, you run a concurrent program named "DQM Staging Program". This process de-normalises TCA data into staging tables. Transformed values for columns are stored in staging table. This will be clear when you go through the PowerPoint presentation.


What are the staging tables?
There are a list staging tables where DQM Transformed data is stored. Each transformed value resides in a different column. The transformed values used for Acquisition are indexed by Oracle.
The list of tables are
HZ_STAGED_CONTACTS
HZ_STAGED_CONTACT_POINTS
HZ_STAGED_PARTIES
HZ_STAGED_PARTY_SITES



What is event "oracle.apps.ar.hz.DQM.realtimesync" which keeps on firing when employee records or the TCA records are modified?
As you know, the TCA data is transformed and de-normalized into staging tables. Any changes to the TCA data must update the DQM staging tables, so that changes can be searched upon in DQM. The subscription of this event reads the pending entries in staging interface table HZ_DQM_SYNC_INTERFACE and transfers them to DQM staging tables after having applied the transformations.

NOTE: Although the event name is realtimesync, Oracle has disabled the realtime option for this event subscription. Hence synchronization with DQM staging tables happens in deferred mode only.


Here is the link to DQM presentation
Link for DQM and TCA presentation 

References:
http://oracle.anilpassi.com/dqm-tca.html