custom algorithm

We have a table containg about 600k entries and growing. Table contains prefix, cost . Prefix will repeat a lot of times, but with a different cost per entry. We need an algorithm designed that compares the entries in the table and produces three lists: 1. a list containing the minimum prices (minimum cost per prefix) 2. a list containing the maximum prices (maximum cost per prefix) 3. a list containing average price per prefix Please understand we already did the following: result = select * from table foreach result do newresult = select * from table where prefix=[login to view URL] if [login to view URL]>newresult... For the 600k queries times 600k queries this is taking too long (hours and hours) We need a faster algorithm. Please do not bid with offers for building a website or other unrelated stuff.