Databases Reference
In-Depth Information
vary in a synchronized fashion through time. Remember that the individ-
ual sales transactions are grained at the time of day down to the second;
meanwhile, the price of gasoline is grained at the date. In that way, trans-
actions that occurred at March 14, 2011, 8:15:03; March 14, 2011, 8:15:27;
and March 14, 2011, 8:16:23 all share the date March 14, 2011, with the
price of gasoline on March 14, 2011. In that way, gasoline sales transac-
tions and the price of gasoline vary together through time, even though
they are grained at different levels of time granularity.
the more they stAy the sAme
Another simple form of Time Variance occurs when the price of gasoline
doesn't change. In the imaginary gas station, the price of gasoline changed
on the morning of March 14. Then gasoline sales transactions occurred
throughout March 14, 15, and 16. Then, the price of gasoline changed on
the morning of March 17, followed by gasoline sales transactions through-
out the days of March 17, 18, and 19. In this example, the price of gasoline
and sales transactions of gasoline still vary together through time. The
price and sales transactions of gasoline on March 14 can still be found
by querying for March 14. The price and sales transactions of gasoline
on March 15 can still be found by querying for March 15. The difference
in this pattern as compared to the previous pattern is that this time the
price of gasoline on March 14 and 15 is the same. But, remember this is
not Price of Gasoline Variance; rather, it is Time Variance. So, the price
of gasoline on a date and the gasoline sales transactions throughout that
date all share the same date. In that way, the price of gasoline and gasoline
sales transactions both vary together through time by synchronizing on
the date.
In this example of a gas station, the data warehouse has three objects.
They are the price of gasoline, gasoline sales transactions, and time (i.e.,
the date). In this gas station data warehouse, time (i.e., the date) is the
join between the price of gasoline and gasoline sales transactions. There
is no other connection between the price of gasoline and gasoline sales
transactions. If you want to know the gasoline sales transactions when
the price of gasoline was XYZ, then you query the “price of gasoline”
table for rows where the price of gasoline is XYZ, then use the date from
the “price of gasoline” table to join to the “gasoline sales transactions”
