164. Maximum Gap【H】【80】【桶排序】【VIP】

Given an unsorted array, find the maximum difference between the successive elements in its sorted form.

Try to solve it in linear time/space.

Return 0 if the array contains less than 2 elements.

You may assume all elements in the array are non-negative integers and fit in the 32-bit signed integer range.

Credits:
Special thanks to @porker2008 for adding this problem and creating all test cases.

Subscribe to see which companies asked this question

注释掉的方法肯定是比较流氓的了。。

重点在于桶排序

这个思想还是很好用的

利用桶排序思想： 
假设有N个元素数组array,最大值为Max,最小值为Min。 
那么最大差值不会大于ceiling[(Maz - Min) / (N - 1)]。

令bucket（桶）的大小len = ceiling[(Maz - Min) / (N - 1)]，则最多会有(Max - Min) / len + 1个桶 
对于数组中的任意整数K，很容易通过算式index = (K - Min) / len找出其桶的位置，然后维护每一个桶的最大值和最小值。

由于同一个桶内的元素之间的差值至多为len - 1，因此最终答案不会从同一个桶中选择。 

对于每一个非空的桶p，找出下一个非空的桶q，则q.min - p.max可能就是备选答案。返回所有这些可能值中的最大值。

class Solution(object):
    def maximumGap(self, nums):

        if len(nums) < 2:
            return 0

        a = min(nums)
        b = max(nums)
        if a == b:
            return 0

        size = math.ceil((b - a) / (len(nums) + 0.0) )
        #print size
        if size == 0:
            size = 1
        if size != int(size):
            size = int(size) + 1
        else:
            size = int(size)
        #print a,b,size,len(nums)

        bucketMin = [sys.maxint] * (len(nums) + 1)
        bucketMax = [0] * (len(nums) + 1)

        for i in nums:
            index = (i - a) / size
            #print index,i,size
            bucketMin[index] = min(bucketMin[index],i)
            bucketMax[index] = max(bucketMax[index],i)

        prv = bucketMax[0]
        res = 0
        for i in xrange(len(nums)):
            if bucketMin[i] == sys.maxint:
                continue
            res = max(bucketMin[i] - prv,res)
            prv = bucketMax[i]

        #print bucketMax,bucketMin

        #print res
        return res

        '''
        nums.sort()
        i = 0
        res = 0
        while i < len(nums) - 1:
            res = max(res,nums[i+1] - nums[i])
            i += 1
        return res
        '''