IS-624-201 CHAPTER 4 – SUMIT

4.2

The snowflake schema and fact constellation are both variants of the
Star schema model, which consists of a fact table and a set of dimension tables. The snowflake schema has normalized dimension tables, whereas the Fact constellation has a set of fact tables that share some common dimension tables. A Star net query model is a query model, which has set of radial lines emanating from a central point and each radial line represents one dimension and each point (called a “footprint”) along the line is consist a level of the dimension, and each step going out from the center represents a concept hierarchy of the dimension. The star net query model, is querying purpose and provides users with a global view of OLAP operations and helps in decision making.

a)
Three classes of schemas popularly used for modeling data warehouses are the star schema, the snowflake schema, and the fact constellations schema.

(b)

(c)

The operations to be performed are:
• Roll-up on time from day to year

• Slice for time=2004

• Roll-up on patient from individual patient to all

(d) To obtain the same list, write an SQL query assuming the data is stored in a relational database with the schema.

Fee (day, month, year, doctor, hospital, patient, count, charge)

4.4
a)

b)
The following OLAP operations should be performed:

•Roll-up on course from course id to department

• Roll-up on student from student id to university

•Dice on course, student with department=“CS” and university = “Big-University”

C) It contain
5^4 = 625 cuboids.

Q 4.5)
a)

b)
The specific OLAP operations to be performed are:

•Roll-up on date from date id to year

• Roll-up on game from game id to all

• Roll-up on location from location id to location name

• Roll-up on spectator from spectator id to status

• Dice with status=“students”, location name=“GM Place”, and year=2004

c) Bitmap indexing is useful for following reasons: * Low cardinality domain * Long location name can be represent by single bit * Vector representation for long cardinality dimension like “date”

4.8
a) Database like telephone billing company have huge and sparse data cube because it has to removed data in every few month because of detailed database and it’s very hard and expensive to maintain that database.
b)
To overcome sparse matrix problem we follow following steps: 1. Partition array based cube into chunks 2. Compress these chunks to remove cells which don’t have valid data 3. Minimize the visit time for cube cells 4. Sort and compute the planes of data cube in ascending order 5. Keep small plane into main memory and compute only one chuck for big plane

c) only compute those cells which have new data instead of whole cube because in incremental update it compute one chunk at a time and recomputed value propagated to higher cuboid.