[英]Finding the Kth Largest element in a Python List using recursion
Given an input list that contains some random unsorted numbers, I am trying to write a program that outputs the kth largest distinct element in that list.给定一个包含一些随机未排序数字的输入列表,我正在尝试编写一个程序来输出该列表中第 k 个最大的不同元素。 For example:例如:
Input:
el = [10,10, 20,30,40, 40]
k = 2
Output: 30 #Since 30 is the second largest distinct element in the list
The following function, takes as input a list, the pivot Index and k and populates list "lesser" with all elements lesser than the pivot and populates another list "greater" with all elements greater than the pivot.以下函数将列表、主元索引和 k 作为输入,并使用小于主元的所有元素填充列表“较小”,并使用大于主元的所有元素填充另一个“更大”列表。
Now, looking at the length of the two list, I can determine if the kth largest element is in the lesser list or the greater list.现在,查看两个列表的长度,我可以确定第 k 个最大的元素是在较小的列表中还是在较大的列表中。 Now I recursively call the same function.现在我递归地调用相同的函数。 However, my program's output is wrong for certain values of k.但是,对于某些 k 值,我的程序的输出是错误的。
def kthLargest(el, pivotIndex, k):
pivot = el[pivotIndex]
lesser = [] #List to store all elements lesser than pivot
greater = [] #Lsit to store all elements greater than pivot
equals = [] #List to store all elements equal to pivot
for x in el:
if x > pivot:
greater.append(x)
elif x < pivot:
lesser.append(x)
else:
equals.append(x)
g = len(greater) #Length of greater list
l = len(lesser)
if(g == k - 1): #If greater list has k-1 elements, that makes the pivot kth largest element
return pivot
elif(g < k):
return kthLargest(lesser, l - 1, k) #If greater list is smaller than k, kth largest element is in lesser list
else:
return kthLargest(greater, g - 1, k) #Else kth largest element is in greater list
Is there any reason you want to use recursion? 你有什么理由想使用递归吗? To find the kth largest element of a list you have to look through the entire list, so the problem is essentially O(n) complexity anyway. 要查找列表中第k个最大元素,您必须查看整个列表,因此问题本质上是O(n)复杂度。
You could do this without recursion like this: 你可以这样做而不会像这样递归:
el = [10, 10, 53, 20, 30, 40, 59, 40]
k = 2
def kth_largest(input_list, k):
# initialize the top_k list to first k elements and sort descending
top_k = input_list[0:k]
top_k.sort(reverse = True)
for i in input_list[k:]:
if i > top_k[-1]:
top_k.pop() # remove the lowest of the top k elements
top_k.append(i) # add the new element
top_k.sort(reverse = True) # re-sort the list
return top_k[-1] # return the kth largest
kth_largest(el, k)
There's an easy way to do this problem using recursion. 使用递归有一种简单的方法来解决这个问题。 I'm just not sure why you need the pivot in the problem description... For example: 我只是不确定为什么你需要问题描述中的枢轴...例如:
def find_kth(k, arr):
if k == 1:
return max(arr)
m = max(arr)
new_arr = list(filter(lambda a: a != m, arr))
return(find_kth(k-1, new_arr))
Here is a simple solution: 这是一个简单的解决方案:
def kthmax(k, list):
if (k == 1):
return max(list)
else:
m = max(list)
return(kthmax(k-1, [x for x in list if x != m]))
kthmax(3,[4, 6, 2, 7, 3, 2, 6, 6])
Output: 4 产量:4
Algorithm: Take the index of max value and convert to zero. 算法:取最大值的索引并转换为零。
def high(arr,n):
for i in range(n+ 1 ):
arr[arr.index(max(arr))] = 0
return max(arr)
high([1,2,3,4,5], 2)
If we can pass in a list or series that is already sorted in descending order, eg 如果我们可以传入已经按降序排序的列表或系列,例如
el.sort_values(ascending=False, inplace = True)
then you can easily find the kth largest (index,value) tuple using just simple slicing of sorted dataframe column and/or list 然后你可以使用简单的排序数据帧列和/或列表的切片轻松找到第k个最大(索引,值)元组
def kth_largest(input_series, k):
new_series = input_series[k-1:len(input_series)]
return (np.argmax(new_series) , np.max(new_series))
el = pd.Series([10, 10, 53, 20, 30, 40, 59, 40])
print el
k = 2
el.sort_values(ascending=False, inplace=True)
print kth_largest(el, 2)
Output: 30
0 10
1 10
2 53
3 20
4 30
5 40
6 59
7 40
dtype: int64
(2, 53)
My way of finding the Kth largest element is...我找到第 K 个最大元素的方法是......
lst=[6,2,3,4,1,5]
print(sorted(lst,reverse=True)[k-1])
在 S Rohith Kumar 的回答之上,如果输入有重复值,那么答案可以是:
print(sorted(set(lst),reverse=True)[k-1])
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.