SEO

Was looking at Kimball SSIS templates and note that when he is updating Fact
tables, he first dumps the original extract into an "extract" table, then in
a 2nd Data Flow area inside the same control flow, pulls that data and lines
it up with the dimension(s).

I think he's only doing it this way for auditing reasons - is this the norm
or are there other reasons to do it this way?

"Joe H" <horton.rj at gmail> wrote in message
news:%23JaQxvurIHA.1872@TK2MSFTNGP04.phx.gbl...
> Was looking at Kimball SSIS templates and note that when he is updating
> Fact tables, he first dumps the original extract into an "extract" table,
> then in a 2nd Data Flow area inside the same control flow, pulls that data
> and lines it up with the dimension(s).
>
> I think he's only doing it this way for auditing reasons - is this the
> norm or are there other reasons to do it this way?

I think Kimball's "extract" table is analagous to what a lot of people
simply call a "staging" area. It's just a temporary storage space, and it's
pretty common. In addition to auditing, a staging area is often used as a
workspace to perform complex calculations and additional validations that
might be done more efficiently on the server in a set-based fashion (SQL)
before you commit the data in its final form.