Recursively find the kth largest int in list of list of int in Python

Question

As a newbie in programming, I am trying to do a function to find the kth largest int in a list of list of ints. I tried list of ints before and it worked.

However, for

Answer 1

some hints:

write a flatten function to convert list hierarchy to a single collection (with unique elements), sort and pick the k'th element.

sample implementation

def flatten(t):
     for e in t:
         if isinstance(e,list):
             yield from flatten(e)
         else:
             yield e



set(flatten([[1,[]], 9, [[1], [3]], [4]]))

{1, 3, 4, 9}

of course, this approach is not the most efficient (perhaps it's the least efficient). You can return the first k elements from each sub level and prune at k'th level at each merge.

Answer 2

I wrote a generalized solution (you specify k) using extra arguments with default values to track both the smallest set of k found so far, and whether this was the top level of the recursion or a sub-level. Sub-levels return the current smallest subset, the top level returns the last entry (which is the k^th smallest value). The smallest set of k is maintained using heapq's nsmallest method.

import heapq

def find_kth_smallest(k, a_list, smallest_k = [], top_level = True):
    l = len(a_list)
    if l > 1:
        l /= 2
        smallest_k = find_kth_smallest(k, a_list[:l], smallest_k, False)
        smallest_k = find_kth_smallest(k, a_list[l:], smallest_k, False)
    elif l < 1:
        return []
    else:
        if isinstance(a_list[0], list):
            smallest_k = find_kth_smallest(k, a_list[0], smallest_k, False)
        else:
            smallest_k.append(a_list[0])
            smallest_k = heapq.nsmallest(k, smallest_k)
    if top_level:
        return smallest_k[-1]
    else:
        return smallest_k

print find_kth_smallest(3, [10, [9, 8, [[]]], [7, 6, [[5], 4, 3]], 2, 1]) # => 3