Archives of the TeradataForum
Message Posted: Wed, 23 Feb 2000 @ 09:22:03 GMT
I think this 'best practice' is inappropriate in the context of data warehousing. A data warehouse process should be able to exploit opportunities to the full and that means taking data in whatever form it is most quickly available. That might well be with delimiters between fields.
Even once an extract and load become routine it is difficult to justify spending developer time doing a tidy up rather than machine time running the load. There is also the problem that a customised extract is more liable to have inaccuracies than a utility generated one, and would need more maintenance.
For us the issue is how to develop and maintain a load of a 500+ fields file into 255-column tables.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 27 Dec 2016|