Page 202 - DCAP603_DATAWARE_HOUSING_AND_DATAMINING

Data Warehousing and Data Mining

notes company, the measure is the number of sold units, which are described by the corresponding
location, time, and item type all dimensions are assumed to uniquely determine the measure,
e.g., for the reseller company, the location, time, producer, and item type provide all necessary
information to determine context of a particular number of sold units.
There are five basic OLAP commands that are used to perform data retrieval from a Data
warehouse.
1. ROLL UP, which is used to navigate to lower levels of details for a given data cube. This
command takes the current data cube (object) and performs a GROUP BY on one of the
dimensions, e.g., given the total number of sold units by month, it can provide sales
summarised by quarter.
2. DRILL DOWN, which is used to navigate to higher levels of detail. This command is the
opposite of ROLL UP, e.g., given the total number of units sold for an entire continent, it
can provide sales in the U.S.A.
3. SLICE, which provides a cut through a given data cube. This command enables users to
focus on some specific slice of data inside the cube, e.g., the user may want to look at the
data concerning unit sales only in Mumbai.
4. DICE, which provides just one cell from the cube (the smallest slice), e.g. it can provide
data concerning the number of sold Canon printers in May 2002 in Lucknow.
5. PIVOT, which rotates the cube to change the perspective, e.g., the “time item” perspective
may be changed into “time location.”

These commands, in terms of their specification and execution, are usually carried out using
a point-and-click interface, and therefore we do not describe their syntax. Instead, we give
examples for each of the above OLAP commands.

ROLL UP Command

The ROLL UP allows the user to summarise data into a more general level in hierarchy. For
instance, if the user currently analyses the number of sold CPU units for each month in the first
half of 2002, this command will allows him/her to aggregate this information into the first two
quarters. Using ROLL UP, the view

figure 10.10: example roLL up command

196 LoveLy professionaL university