WAREHOUSING:

If we wish to make only incremental changes to a warehouse every night, what
sort of support do we need from each data source?  Name 3 options.

Suppose I have a fact table with attributes: itemID, date, location, and
sales volume.  How many dimensions do I have?  Is this schema a star,
snowflake, neither, both, or not specified.?

Given a query:
SELECT A,B,C,D,SUM(Sales)
FROM Rel
GROUPBY B, ROLLUP C,D, CUBE A;
Explain what aggregates will be computed.
How many columns will the resulting relation have?

Consider a university that has a STUDENT relation, with one row for each of
10,000 students.  Each student tuple is 512 bytes in size.  Among others,
attributes include GENDER (50% Male, 50% Female), SCHOLARSHIP (10% YES,
90% NO), and GPA (stored as a 4 Byte real number).
I want to find out the average GPA of all female scholarship students.
My machine has pages that are only 2KBytes, and my query is only allocated 8
pages of memory.  Compute the (IO) cost of evaluating my query in each of the
following cases:
a. You have 2 individual attribute hash indices on GENDER and SCHOLARSHIP.
b. You have 2 individual attribute bitmap indices on GENDER and SCHOLARHIP.
c. You have a joint, 2-attribute, hash index on these two attributes.

Compare the storage required for a bitslice index versus a B-tree index for
some attribute that takes integer values between 0 and 1000,000.
Assume any additional information you need regarding the relation in
question.