I am a believer in a clearly defined staging process. In the situation you describe I see Inbound staging libraries, Processing staging libraries, and then of course Production staging libraries.

Each of those stages in the process have differing requirements for security, integrity, disaster recovery, user interaction and no doubt other issues.

In my opinion.


On 1/28/2014 3:48 PM, Charles Wilt wrote:
Another philosophical question for you all..

I'm beginning to work with a once a week incremental data feed from an
outside source that consists of about 30 tables.

I'll end up with a set of 30 production tables and 30 staging tables.

I intend to have the production tables in a separate library from the rest
of my application (currently in 4 data libraries)

I'm trying to decide rather of not to keep the staging tables in the same
library as the production tables or in another library.

Obviously, if I keep them in the same library, I'd need need to have the
staging table named differently from the production version. Whereas if I
put the staging tables in another library I could have the same name.

I was initially leaning toward same name, different libraries. But now I'm
thinking I want different names so I can use unqualified names and depend
on the library list resolution. Primarily so I can have TEST/DEV
environments within my change management system (Aldon)

But that leaves me with deciding rather or not to have the tables in one
library or two.

Thoughts?

Charles





Return to Archive home page | Return to MIDRANGE.COM home page