Databases Reference
In-Depth Information
units every minute, an enterprise will not change the product, price, pro-
motion, or placement thousands of times a minute. Typically, the shortest
periodicity for the dimensions is the day. A gas station typically changes
the price at the beginning of the day. A retail store may change the price at
the beginning and ending of every promotion.
The periodicity of transactions is typically short. The periodicity of
dimensions is typically long. The goal of a data warehouse is to rational-
ize the two forms of periodicity into one representation of the enterprise.
This means that all the retail sales transactions of gasoline during the day
of March 14 and the price of gasoline on March 14 should both have an
attribute that shows they both happened on March 14. The individual sales
transactions also show the time of day grained at the second and the date,
which is March 14. The price dimension of that gasoline shows the price
of gasoline on March 14. An integrated view of the price and sales data
of that gas station would show the date and time of each individual sales
transaction and the price of gasoline on that same day.
the more thINGs chANGe
In the integrated view of an enterprise that just happens to be a gas sta-
tion, sales transactions throughout the day of March 14 and the price of
gasoline on March 14 all indicate they occurred on March 14. If the price
of gasoline in the gas station changes on March 15, then for that gasoline
price and all the sales transactions that occurred throughout the day, the
data warehouse would indicate they all occurred on March 15. This same
pattern can repeat on March 16, 17, 18…ad infinitum. Each morning the
price changes and then throughout that day the sales transactions occur,
all of them indicating that day's date as the day on which they occurred. As
they change every day, they stay together every day. On every day, the price
of gasoline and sales transactions indicate the same date. Then, if you want
to know the price of gasoline and the sales of gasoline on March 14, you
could query the data warehouse for the date of March 14. If you want to
know the price of gasoline and the sales of gasoline on March 15, you query
the data warehouse for the date of March 15. The same is true for March 16,
17, 18…ad infinitum. That, in its simplest form, is Time Variance.
In this gas station data warehouse the price of gasoline and sale transac-
tions of gasoline are synchronized by date, that is, Time Variance. They
