Hello everyone,
i have an assignment and i would really appreciate it if any of you could help me.

The assignment is:
"You have a relation EMPLOYEES (clustered) sized B(R)>M,where M being the memory of your system.The EMPLOYEES has one attribute salary and you want to calculate the median of the salary of the employees in this relation:

SELECT MEDIAN(Employees.salary)FROM Employees.

Describe a sufficient algorithm (not SQL/code) that calculates questions as the above.
What is the cost of your algorithm"

I'd be very thankful if someone could help.

What sort of help would you like? It would be a lot easier to help you if you could ask us a specific question about your assignment instead of just posting the assignement verbatim.

-PatP

I'm not sure whether the cost is going to be 2,5B(R) or 3B(R).
I mean what do you do:read from the disk and sort in the memory and then flush to disk which has cost 2B(R) and then just read by the half and you'll find the median OR write to memory once more to sort the sorted sublists and then just pull the middle tuple therefor the cost is 3B(R) plus one tuplet???

No problem.

First, get some more M. M is cheap. No excuse for running low on M these days.

Count the number of employees. Divide by two. Step through the employees sorted by salary until you hit that number.

$29.99, and I accept PayPal.

Again, no problem.

Here are some more ways:
Selection algorithm - Wikipedia, the free encyclopedia