Data integration is the process of combining multiple data sets into a single hopefully coherent data set.  Prior to Tandem Primary Keys ^tm, data integration required that data from multiple sources be combined and cleansed prior to storage in a target database.  The expectation was that all data be extracted and transformed into pristine, high quality data before it is loaded into the target database.  Then, once the target database has been loaded, the data will remain pristine and consistent with volumes of new data being added often on a daily basis.

This one-shot data integration/cleansing process is an unrealistic expectation.  Even the use of expensive, sophisticated, ETL tools combined with data cleansing tools can not guarantee the proper level of data quality.  A data steward’s interpretations of data inconsistencies and conflicts can be helpful but is usually an expensive alternative.  In any event, data that change with time (mergers, divestitures, relocations) can not be anticipated and cannot be adequately represented in most relational databases.  The resulting data integration quality is often less than what is required. 

With Tandem Primary Keys, the data integration process may be accomplished using many different methods including any of the methods commonly used today.  Since we added a level of isolation to our data, it is now possible to integrate, cleanse and enrich the data without removing it from the source database!  The business data may simply be recast with a set of more universal reference data.  This recasting of business data does not need to change the data context  of the original data in any way.

Tandem Primary Keys ^tm is a trademark of Strategic Insights, Inc.  Tandem Primary Keys is patented by U.S. Patent No. 6801915.    Copyright ©2005 Strategic Insights, Inc.  All Rights Reserved.