Database Reference
In-Depth Information
examined based on street addresses. Cities fall somewhere in between the two, but if you look closely, you
see that cities are never repeated. Therefore, it is unlikely that useful information can be extracted regarding
sales based on the city in which the stores are located.
One question we forgot to ask during the interview process was how long the client has been collecting
this particular set of data. If you went back and asked this question, you would find that the number
of stores, publishers, and authors seldom changed over the last few years. If a lot of change had been
occurring, we might have decided that the city column would become more useful as time went on.
However, because change is slow or nonexistent, all the city column will do is provide additional data
in the data warehouse without much additional information. using the “keep it simple” rule, you might
choose to exclude the street address, city, and zip code from this first version of the data warehouse. As
time goes by, if you decide that this data becomes useful, you can add it during the creation of future
versions.
Note:
We are leaving each store's state data out of the first version of the design. As we see later in the
book, we will come to regret this decision. But, it allows us a chance to examine what to do if something
is missing from an initial version.
3.
Close the query window by clicking the
x
on this window's tab.
Reviewing the Data in the Pub_Info Table
A number of columns in various supporting tables are not necessarily useful for this version but could be
useful in the future. For example, the publishers' information table, or pub_info as it is called, has little in it
that is of use to us at this time (Figure
3-16
).
