Now
I am working as a data modeler on a project. As soon as I joined the
project, I read the ISP summary for the project. And the ISP report
recommended that we should change a data model with regards to storing
historical data.
The following was the ERD in the ISP summary:
* AS-IS data model
There are around 800 base tables which have additional tables for storing historical data. The ISP report said that we should merge about 600 tables for historical data into one table.
The
thing is that application designers already decided to adopt the TO-BE
model before a data modeler was assigned to this project. And the data
modeler who was assigned later to the subject area objected to using the
TO-BE model, but the PMO rejected the modeler's offer. I was mercifully
assigned to another subject area.
What is the use of merging tables? Does it matter to have a lot of tables for historical data?
--Due to the volume limit of this bulletin board I put up the rest of this article in the next post.
