Problem: The Source-System added a new descriptive attribute to a source-table.
We need to include the new data in our Raw-Data-Vault(-Satellite)-World.
We have two possibilities:
- Add a new Hub-Satellite which contains besides the metadata only the new attribute from the time on the attribute appeared.
- Stop filling the currrent Hub-Satellite and fill a new Satellite which is a structural copy from the old Satellite plus the new attribute from the time on the attribute appeared
Case 1.: I think this is the normal case. The existing dataflow based on the data without the new Attribute is untouched and makes still sense for the business so there is no need to change or anounce anything concerning the dataflow or any report based on the information without the new Attribute.
We can develope an extended dataflow containing the new Attribute and a new report which is very similar to the ‘old’ report without the new Attribute.
After successfully completion we can set the old report to ‘deprecated’, if there is no need to use the old report because the new report, containing the new Attribute, covers the old situation, too. After a given period of time the business should use only the new report and the old one can be removed.
Case 2.: These situation is a complete different one: The old satellite will be not filled anymore after appearance of the new attribute. So the old dataflow and reports based on the old satellite shows data only till the arrival of the new attribute-data.
This could be necessary if the new attriute is a game-changer concerning the old characteristic of the data so that the old report does not make sense anymore by knowing the new attribute.
So the old report shows only the data before and a an new report shows only the data from the date on the new attribute-values are available.
It is of course possible to connect the old an the new satellite by a view for example, but here we have the problem of the values for the new attribute in the history…
Any thought on this topic is appreciated!