Develop Reference

How can I partition `itertools.combinations` such that I can process the results in parallel?

I have a massive quantity of combinations (86 choose 10, which yields 3.5 trillion results) and I have written an algorithm which is capable of processing 500,000 combinations per second. I would not like to wait 81 days to see the final results, so naturally I am inclined to separate this into many processes to be handled by my many cores.
Consider this naive approach:
import itertools
from concurrent.futures import ProcessPoolExecutor
def algorithm(combination):
# returns a boolean in roughly 1/500000th of a second on average
def process(combinations):
for combination in combinations:
if algorithm(combination):
# will be very rare (a few hundred times out of trillions) if that matters
print("Found matching combination!", combination)
combination_generator = itertools.combinations(eighty_six_elements, 10)
# My system will have 64 cores and 128 GiB of memory
with ProcessPoolExecutor(workers=63) as executor:
# assign 1,000,000 combinations to each process
# it may be more performant to use larger batches (to avoid process startup overhead)
# but eventually I need to start worrying about running out of memory
group = []
for combination in combination_generator:
group.append(combination)
if len(group) >= 1_000_000:
executor.submit(process, group)
group = []
This code "works", but it has virtually no performance gain over a single-threaded approach, since it is bottlenecked by the generation of the combinations for combination in combination_generator.
How can I pass this computation off to the child-processes so that it can be parallelized? How can each process generate a specific subset of itertools.combinations?
p.s. I found this answer, but it only deals with generating single specified elements, whereas I need to efficiently generate millions of specified elements.

I'm the author of one answer to the question you already found for generating the combination at a given index. I'd start with that: Compute the total number of combinations, divide that by the number of equally sized subsets you want, then compute the cut-over for each of them. Then do your subprocess tasks with these combinations as bounds. Within each subprocess you'd do the iteration yourself, not using itertools. It's not hard:
def next_combination(n: int, c: list[int]):
"""Compute next combination, in lexicographical order.
Args:
n: the number of items to choose from.
c: a list of integers in strictly ascending order,
each of them between 0 (inclusive) and n (exclusive).
It will get modified by the call.
Returns: the list c after modification,
or None if this was the last combination.
"""
i = len(c)
while i > 0:
i -= 1
n -= 1
if c[i] == n: continue
c[i] += 1
for j in range(i + 1, len(c)):
c[j] = c[j - 1] + 1
return c
return None
Note that both the code above and the one from my other answer assume that you are looking for combinations of elements from range(n). If you want to combine other elements, do combinations for this range then use the elements of the found combinations as indices into your sequence of actual things you want to combine.
The main advantage of the approach above is that it ensures equal batch size, which might be useful if processing time is expected to be mostly determined by batch size. If processing time still varies greatly even for batches of the same size, that might be too much effort. I'll post an alternative answer addressing that.

You can do a recursive divide-and-conquer approach, where you make a decision based on the expected number of combinations. If it is small, use itertools. If it is large, handle the case of the first element being included and of it being excluded both in recursive calls.
The result does not ensure batches of equal size, but it does give you an upper bound on the size of each batch. If processing time of each batch is somewhat varied anyway, that might be good enough.
T = typing.TypeVar('T')
def combination_batches(
seq: collections.abc.Sequence[T],
r: int,
max_batch_size: int,
prefix: tuple[T, ...] = ()
) -> collections.abc.Iterator[collections.abc.Iterator[tuple[T, ...]]]:
"""Compute batches of combinations.
Each yielded value is itself a generator over some of the combinations.
Taken together they produce all the combinations.
Args:
seq: The sequence of elements to choose from.
r: The number of elements to include in each combination.
max_batch_size: How many elements each returned iterator
is allowed to iterate over.
prefix: Used during recursive calls, prepended to each returned tuple.
Yields: generators which each generate a subset of all the combinations,
in a way that generators together yield every combination exactly once.
"""
if math.comb(len(seq), r) > max_batch_size:
# One option: first element taken.
yield from combination_batches(
seq[1:], r - 1, max_batch_size, prefix + (seq[0],))
# Other option: first element not taken.
yield from combination_batches(
seq[1:], r, max_batch_size, prefix)
return
yield (prefix + i for i in itertools.combinations(seq, r))
See https://ideone.com/GD6WYl for a more compete demonstration.
Note that I don't know how well the process pool executor deals with generators as arguments, whether it is able to just forward a short description of each. There is a chance that in order to ship the generator to a subprocess it will actually generate all the values. So instead of yielding the generator expression the way I did, you might want to yield some object which pickles more nicely but still offers iteration over the same values.

how to calculate the minimum unfairness sum of a list

I have tried to summarize the problem statement something like this::
Given n, k and an array(a list) arr where n = len(arr) and k is an integer in set (1, n) inclusive.
For an array (or list) myList, The Unfairness Sum is defined as the sum of the absolute differences between all possible pairs (combinations with 2 elements each) in myList.
To explain: if mylist = [1, 2, 5, 5, 6] then Minimum unfairness sum or MUS. Please note that elements are considered unique by their index in list not their values
MUS = |1-2| + |1-5| + |1-5| + |1-6| + |2-5| + |2-5| + |2-6| + |5-5| + |5-6| + |5-6|
If you actually need to look at the problem statement, It's HERE
My Objective
given n, k, arr(as described above), find the Minimum Unfairness Sum out of all of the unfairness sums of sub arrays possible with a constraint that each len(sub array) = k [which is a good thing to make our lives easy, I believe :) ]
what I have tried
well, there is a lot to be added in here, so I'll try to be as short as I can.
My First approach was this where i used itertools.combinations to get all the possible combinations and statistics.variance to check its spread of data (yeah, I know I'm a mess).
Before you see the code below, Do you think these variance and unfairness sum are perfectly related (i know they are strongly related) i.e. the sub array with minimum variance has to be the sub array with MUS??
You only have to check the LetMeDoIt(n, k, arr) function. If you need MCVE, check the second code snippet below.
from itertools import combinations as cmb
from statistics import variance as varn
def LetMeDoIt(n, k, arr):
v = []
s = []
subs = [list(x) for x in list(cmb(arr, k))] # getting all sub arrays from arr in a list
i = 0
for sub in subs:
if i != 0:
var = varn(sub) # the variance thingy
if float(var) < float(min(v)):
v.remove(v[0])
v.append(var)
s.remove(s[0])
s.append(sub)
else:
pass
elif i == 0:
var = varn(sub)
v.append(var)
s.append(sub)
i = 1
final = []
f = list(cmb(s[0], 2)) # getting list of all pairs (after determining sub array with least MUS)
for r in f:
final.append(abs(r[0]-r[1])) # calculating the MUS in my messy way
return sum(final)
The above code works fine for n<30 but raised a MemoryError beyond that.
In Python chat, Kevin suggested me to try generator which is memory efficient (it really is), but as generator also generates those combination on the fly as we iterate over them, it was supposed to take over 140 hours (:/) for n=50, k=8 as estimated.
I posted the same as a question on SO HERE (you might wanna have a look to understand me properly - it has discussions and an answer by fusion which takes me to my second approach - a better one(i should say fusion's approach xD)).
Second Approach
from itertools import combinations as cmb
def myvar(arr): # a function to calculate variance
l = len(arr)
m = sum(arr)/l
return sum((i-m)**2 for i in arr)/l
def LetMeDoIt(n, k, arr):
sorted_list = sorted(arr) # i think sorting the array makes it easy to get the sub array with MUS quickly
variance = None
min_variance_sub = None
for i in range(n - k + 1):
sub = sorted_list[i:i+k]
var = myvar(sub)
if variance is None or var<variance:
variance = var
min_variance_sub=sub
final = []
f = list(cmb(min_variance_sub, 2)) # again getting all possible pairs in my messy way
for r in f:
final.append(abs(r[0] - r[1]))
return sum(final)
def MainApp():
n = int(input())
k = int(input())
arr = list(int(input()) for _ in range(n))
result = LetMeDoIt(n, k, arr)
print(result)
if __name__ == '__main__':
MainApp()
This code works perfect for n up to 1000 (maybe more), but terminates due to time out (5 seconds is the limit on online judge :/ ) for n beyond 10000 (the biggest test case has n=100000).
=====
How would you approach this problem to take care of all the test cases in given time limits (5 sec) ? (problem was listed under algorithm & dynamic programming)
(for your references you can have a look on
successful submissions(py3, py2, C++, java) on this problem by other candidates - so that you can
explain that approach for me and future visitors)
an editorial by the problem setter explaining how to approach the question
a solution code by problem setter himself (py2, C++).
Input data (test cases) and expected output
Edit1 ::
For future visitors of this question, the conclusions I have till now are,
that variance and unfairness sum are not perfectly related (they are strongly related) which implies that among a lots of lists of integers, a list with minimum variance doesn't always have to be the list with minimum unfairness sum. If you want to know why, I actually asked that as a separate question on math stack exchange HERE where one of the mathematicians proved it for me xD (and it's worth taking a look, 'cause it was unexpected)
As far as the question is concerned overall, you can read answers by archer & Attersson below (still trying to figure out a naive approach to carry this out - it shouldn't be far by now though)
Thank you for any help or suggestions :)

You must work on your list SORTED and check only sublists with consecutive elements. This is because BY DEFAULT, any sublist that includes at least one element that is not consecutive, will have higher unfairness sum.
For example if the list is
[1,3,7,10,20,35,100,250,2000,5000] and you want to check for sublists with length 3, then solution must be one of [1,3,7] [3,7,10] [7,10,20] etc
Any other sublist eg [1,3,10] will have higher unfairness sum because 10>7 therefore all its differences with rest of elements will be larger than 7
The same for [1,7,10] (non consecutive on the left side) as 1<3
Given that, you only have to check for consecutive sublists of length k which reduces the execution time significantly
Regarding coding, something like this should work:
def myvar(array):
return sum([abs(i[0]-i[1]) for i in itertools.combinations(array,2)])
def minsum(n, k, arr):
res=1000000000000000000000 #alternatively make it equal with first subarray
for i in range(n-k):
res=min(res, myvar(l[i:i+k]))
return res

I see this question still has no complete answer. I will write a track of a correct algorithm which will pass the judge. I will not write the code in order to respect the purpose of the Hackerrank challenge. Since we have working solutions.
The original array must be sorted. This has a complexity of O(NlogN)
At this point you can check consecutive sub arrays as non-consecutive ones will result in a worse (or equal, but not better) "unfairness sum". This is also explained in archer's answer
The last check passage, to find the minimum "unfairness sum" can be done in O(N). You need to calculate the US for every consecutive k-long subarray. The mistake is recalculating this for every step, done in O(k), which brings the complexity of this passage to O(k*N). It can be done in O(1) as the editorial you posted shows, including mathematic formulae. It requires a previous initialization of a cumulative array after step 1 (done in O(N) with space complexity O(N) too).
It works but terminates due to time out for n<=10000.
(from comments on archer's question)
To explain step 3, think about k = 100. You are scrolling the N-long array and the first iteration, you must calculate the US for the sub array from element 0 to 99 as usual, requiring 100 passages. The next step needs you to calculate the same for a sub array that only differs from the previous by 1 element 1 to 100. Then 2 to 101, etc.
If it helps, think of it like a snake. One block is removed and one is added.
There is no need to perform the whole O(k) scrolling. Just figure the maths as explained in the editorial and you will do it in O(1).
So the final complexity will asymptotically be O(NlogN) due to the first sort.

Avoid generating duplicate values from random

I want to generate random numbers and store them in a list as the following:
alist = [random.randint(0, 2 ** mypower - 1) for _ in range(total)]
My concern is the following: I want to generate total=40 million values in the range of (0, 2 ** mypower - 1). If mypower = 64, then alist will be of size ~20GB (40M*64*8) which is very large for my laptop memory. I have an idea to iteratively generate chunk of values, say 5 million at a time, and save them to a file so that I don't have to generate all 40M values at once. My concern is that if I do that in a loop, it is guaranteed that random.randint(0, 2 ** mypower - 1) will not generate values that were already generated from the previous iteration? Something like this:
for i in range(num_of_chunks):
alist = [random.randint(0, 2 ** mypower - 1) for _ in range(chunk)]
# save to file

Well, since efficiency/speed doesn't matter, I think this will work:
s = set()
while len(s) < total:
s.add(random.randint(0, 2 ** mypower - 1))
alist = list(s)
Since sets can only have unique elements in it, i think this will work well enough

To guarantee unique values you should avoid using random. Instead you should use an encryption. Because encryption is reversible, unique inputs guarantee unique outputs, given the same key. Encrypt the numbers 0, 1, 2, 3, ... and you will get guaranteed unique random-seeming outputs back providing you use a secure encryption. Good encryption is designed to give random-seeming output.
Keep track of the key (essential) and how far you have got. For your first batch encrypt integers 0..5,000,000. For the second batch encrypt 5,000,001..10,000,000 and so on.
You want 64 bit numbers, so use DES in ECB mode. DES is a 64-bit cipher, so the output from each encryption will be 64 bits. ECB mode does have a weakness, but that only applies with identical inputs. You are supplying unique inputs so the weakness is not relevant for your particular application.
If you need to regenerate the same numbers, just re-encrypt them with the same key. If you need a different set of random numbers (which will duplicate some from the first set) then use a different key. The guarantee of uniqueness only applies with a fixed key.

One way to generate random values that don't repeat is first to create a list of contiguous values
l = list(range(1000))
then shuffle it:
import random
random.shuffle(l)
You could do that several times, and save it in a file, but you'll have limited ranges since you'll never see the whole picture because of your limited memory (it's like trying to sort a big list without having the memory for it)
As someone noted, to get a wide span of random numbers, you'll need a lot of memory, so simple but not so efficient.
Another hack I just though of: do the same as above but generate a range using a step. Then in a second pass, add a random offset to the values. Even if the offset values repeat, it's guaranteed to never generate the same number twice:
import random
step = 10
l = list(range(0,1000-step,step))
random.shuffle(l)
newlist = [x+random.randrange(0,step) for x in l]
with the required max value and number of iterations that gives:
import random
number_of_iterations = 40*10**6
max_number = 2**64
step = max_number//number_of_iterations
l = list(range(0,max_number-step,step))
random.shuffle(l)
newlist = [x+random.randrange(0,step) for x in l]
print(len(newlist),len(set(newlist)))
runs in 1-2 minutes on my laptop, and gives 40000000 distinct values (evenly scattered across the range)

Usually random number generators are not really random at all. In fact, this is quite helpful in some situations. If you want the values to be unique after the second iteration, send it a different seed value.
random.seed()
The same seed will generate the same list, so if you want the next iteration to be the same, use the same seed. if you want it to be different, use a different seed.

It may need High CPU and Physical memory!
I suggest you to classify your data.
For Example you can Save:
All numbers starting with 10 and are 5 characters(Example:10365) to 10-5.txt
All numbers starting with 11 and are 6 characters(Example:114567) to 11-6.txt
Then for checking a new number:
For Example my number is 9256547
It starts with 92 and is 7 characters.
So I search 92-7.txt for this number and if it isn't duplicate,I will add it to 92-7.txt
Finally,You can join all files together.
Sorry If I have mistakes.My main language isn't English.

Algos - Delete Extremes From A List of Integers in Python?

I want to eliminate extremes from a list of integers in Python. I'd say that my problem is one of design. Here's what I cooked up so far:
listToTest = [120,130,140,160,200]
def function(l):
length = len(l)
for x in xrange(0,length - 1):
if l[x] < (l[x+1] - l[x]) * 4:
l.remove(l[x+1])
return l
print function(listToTest)
So the output of this should be: 120,130,140,160 without 200, since that's way too far ahead from the others.
And this works, given 200 is the last one or there's only one extreme. Though, it gets problematic with a list like this:
listToTest = [120,200,130,140,160,200]
Or
listToTest = [120,130,140,160,200,140,130,120,200]
So, the output for the last list should be: 120,130,140,160,140,130,120. 200 should be gone, since it's a lot bigger than the "usual", which revolved around ~130-140.
To illustrate it, here's an image:
Obviously, my method doesn't work. Some thoughts:
- I need to somehow do a comparison between x and x+1, see if the next two pairs have a bigger difference than the last pair, then if it does, the pair that has a bigger difference should have one element eliminated (the biggest one), then, recursively do this again. I think I should also have an "acceptable difference", so it knows when the difference is acceptable and not break the recursivity so I end up with only 2 values.
I tried writting it, but no luck so far.

You can use statistics here, eliminating values that fall beyond n standard deviations from the mean:
import numpy as np
test = [120,130,140,160,200,140,130,120,200]
n = 1
output = [x for x in test if abs(x - np.mean(test)) < np.std(test) * n]
# output is [120, 130, 140, 160, 140, 130, 120]

Your problem statement is not clear. If you simply want to remove the max and min then that is a simple
O(N) with 2 extra memory- which is O(1)
operation. This is achieved by retaining the current min/max value and comparing it to each entry in the list in turn.
If you want the min/max K items it is still
O(N + KlogK) with O(k) extra memory
operation. This is achieved by two priorityqueue's of size K: one for the mins, one for the max's.
Or did you intend a different output/outcome from your algorithm?
UPDATE the OP has updated the question: it appears they want a moving (/windowed) average and to delete outliers.
The following is an online algorithm -i.e. it can handle streaming data http://en.wikipedia.org/wiki/Online_algorithm
We can retain a moving average: let's say you keep K entries for the average.
Then create a linked list of size K and a pointer to the head and tail. Now: handling items within the first K entries needs to be thought out separately. After the first K retained items the algo can proceed as follows:
check the next item in the input list against the running k-average. If the value exceeds the acceptable ratio threshold then put its list index into a separate "deletion queue" list. Otherwise: update the running windowed sum as follows:
(a) remove the head entry from the linked list and subtract its value from the running sum
(b) add the latest list entry as the tail of the linked list and add its value to the running sum
(c) recalculate the running average as the running sum /K
Now: how to handle the first K entries? - i.e. before we have a properly initialized running sum?
You will need to make some hard-coded decisions here. A possibility:
run through all first K+2D (D << K) entries.
Keep d max/min values
Remove the d (<< K) max/min values from that list

Finding Nth item of unsorted list without sorting the list

Hey. I have a very large array and I want to find the Nth largest value. Trivially I can sort the array and then take the Nth element but I'm only interested in one element so there's probably a better way than sorting the entire array...

A heap is the best data structure for this operation and Python has an excellent built-in library to do just this, called heapq.
import heapq
def nth_largest(n, iter):
return heapq.nlargest(n, iter)[-1]
Example Usage:
>>> import random
>>> iter = [random.randint(0,1000) for i in range(100)]
>>> n = 10
>>> nth_largest(n, iter)
920
Confirm result by sorting:
>>> list(sorted(iter))[-10]
920

Sorting would require O(nlogn) runtime at minimum - There are very efficient selection algorithms which can solve your problem in linear time.
Partition-based selection (sometimes Quick select), which is based on the idea of quicksort (recursive partitioning), is a good solution (see link for pseudocode + Another example).

A simple modified quicksort works very well in practice. It has average running time proportional to N (though worst case bad luck running time is O(N^2)).
Proceed like a quicksort. Pick a pivot value randomly, then stream through your values and see if they are above or below that pivot value and put them into two bins based on that comparison.
In quicksort you'd then recursively sort each of those two bins. But for the N-th highest value computation, you only need to sort ONE of the bins.. the population of each bin tells you which bin holds your n-th highest value. So for example if you want the 125th highest value, and you sort into two bins which have 75 in the "high" bin and 150 in the "low" bin, you can ignore the high bin and just proceed to finding the 125-75=50th highest value in the low bin alone.

You can iterate the entire sequence maintaining a list of the 5 largest values you find (this will be O(n)). That being said I think it would just be simpler to sort the list.

You could try the Median of Medians method - it's speed is O(N).

Use heapsort. It only partially orders the list until you draw the elements out.

You essentially want to produce a "top-N" list and select the one at the end of that list.
So you can scan the array once and insert into an empty list when the largeArray item is greater than the last item of your top-N list, then drop the last item.
After you finish scanning, pick the last item in your top-N list.
An example for ints and N = 5:
int[] top5 = new int[5]();
top5[0] = top5[1] = top5[2] = top5[3] = top5[4] = 0x80000000; // or your min value
for(int i = 0; i < largeArray.length; i++) {
if(largeArray[i] > top5[4]) {
// insert into top5:
top5[4] = largeArray[i];
// resort:
quickSort(top5);
}
}

As people have said, you can walk the list once keeping track of K largest values. If K is large this algorithm will be close to O(n2).
However, you can store your Kth largest values as a binary tree and the operation becomes O(n log k).
According to Wikipedia, this is the best selection algorithm:
function findFirstK(list, left, right, k)
if right > left
select pivotIndex between left and right
pivotNewIndex := partition(list, left, right, pivotIndex)
if pivotNewIndex > k // new condition
findFirstK(list, left, pivotNewIndex-1, k)
if pivotNewIndex < k
findFirstK(list, pivotNewIndex+1, right, k)
Its complexity is O(n)

One thing you should do if this is in production code is test with samples of your data.
For example, you might consider 1000 or 10000 elements 'large' arrays, and code up a quickselect method from a recipe.
The compiled nature of sorted, and its somewhat hidden and constantly evolving optimizations, make it faster than a python written quickselect method on small to medium sized datasets (< 1,000,000 elements). Also, you might find as you increase the size of the array beyond that amount, memory is more efficiently handled in native code, and the benefit continues.
So, even if quickselect is O(n) vs sorted's O(nlogn), that doesn't take into account how many actual machine code instructions processing each n elements will take, any impacts on pipelining, uses of processor caches and other things the creators and maintainers of sorted will bake into the python code.

You can keep two different counts for each element -- the number of elements bigger than the element, and the number of elements lesser than the element.
Then do a if check N == number of elements bigger than each element
-- the element satisfies this above condition is your output
check below solution
def NthHighest(l,n):
if len(l) <n:
return 0
for i in range(len(l)):
low_count = 0
up_count = 0
for j in range(len(l)):
if l[j] > l[i]:
up_count = up_count + 1
else:
low_count = low_count + 1
# print(l[i],low_count, up_count)
if up_count == n-1:
#print(l[i])
return l[i]
# # find the 4th largest number
l = [1,3,4,9,5,15,5,13,19,27,22]
print(NthHighest(l,4))
-- using the above solution you can find both - Nth highest as well as Nth Lowest

If you do not mind using pandas then:
import pandas as pd
N = 10
column_name = 0
pd.DataFrame(your_array).nlargest(N, column_name)
The above code will show you the N largest values along with the index position of each value.
Hope it helps. :-)
Pandas Nlargest Documentation