Welcome to TDBSoverflow, Our class's own StackOverflow. Our rules:
  1. Use only meaningful and self-explanatory titles
  2. Tag your questions with meaningful keywords
  3. Use upvotes and downvotes to rate the answers
  4. When you receive a satisfying answer - Click the "V" button
Remember: you may get up to 5 bonus points to your final grade!

Changing hash function in Hash-Join

+2 votes

In the Hash-Join algorithm we learned in lecture 13 (Query Execution- slide 25), we change the hash function in case one of the partitions is bigger than M-2 (when M is the memory size).

My question is, if the partition reached M-2 in size, it might be because all the attributes hashed are equal and hence would result in the same partition for every hash function we choose.

Is this an edge-case which we should disregard for its likelihood or am I missing something here?



asked Feb 3, 2018 by odednavon (350 points)
Good point!
I think it's an edge case you don't need to consider
(In the past exams I saw, they always tell you to assume hash join is performed without the need to repeat the partitioning to buckets).
But better confirm with Amit.

Please log in or register to answer this question.