Comments on: Computer Algorithms: Bucket Sort

By: xyz

xyz — Tue, 02 May 2017 10:19:56 +0000

How did you decide on how many buckets to use ? And how do u calculate which array element goes into which bucket ? If we go by your second example of using ranges to represent bucket, don’t you think there is an overhead in calculating which bucket an element belongs to ..

By: Ed Fancher

Ed Fancher — Wed, 01 Mar 2017 13:48:10 +0000

For the why question above, because most sorting algorithms are at best O(n log n) and at worst O(n^2) or worse. Bucket sort is O(n) (Assuming n is somewhat large and you’re buckets are relatively small as mentioned in the article) That’s the worst time under those assumptions. http://stackoverflow.com/questions/7311415/how-is-the-complexity-of-bucket-sort-is-onk-if-we-implement-buckets-using-lin
For an array with 10,000 elements, that’s a difference between 10,000 and ~100,000 in the best case and between 10,000 and 100,000,000 in the worst case.

By: Ana

Ana — Sun, 08 Feb 2015 19:55:41 +0000

How do you count number of swaps and comparison in Bucket sort? Do you count swaps and comparisons from Insertion sort only or something more?

By: programmer

programmer — Sat, 17 Jan 2015 07:37:48 +0000

why we use bucket sort? what problem is encountered in other sorting algorithms that we need to use bucket sort?…problem statement for bucket sort??

By: thoroc

thoroc — Fri, 08 Aug 2014 14:58:42 +0000

Hi,

I have found your explanation quite useful and I have tried to apply them to test the different sort presented in your blog. The result (in golang) can be found at: https://github.com/thoroc/go_test/blob/master/sort.go

By: APR

APR — Thu, 02 May 2013 01:27:46 +0000

how do we decide on the interval size here? in your second example we could have decided to choose buckets of size (0-5,5-10,10-15,15-20,20-25). What is the best approach to decide bucket size?