Find subsequence of length k with the largest product - arrays

Language: Python
Given an array of integers, return the subsequence of length k which has the largest possible product. If there is more than one valid subsequence that gives the same product, return the one with the largest sum of numbers.
Example 1
array = [-10, -3, 5, 6, -2]
k = 2
Output should be [-10, -3] ( (-10) * (-3) = 30, which is the largest product of the given numbers)
Example 2
array = [10, 3, 5, 6, 20]
k = 3
Output should be [6, 10, 20], (6 * 10 * 20 = 1200)
Example 3
array = [1, -4, 3, -6, 7, 0]
k = 4
Output should be [-6, -4, 3, 7] ( (-6) * (-4) * 3 * 7 = 504)
I've already tried the code
def find_k_prod(arr, k):
arr = sorted(arr)
current_prod = 1
n = len(arr)
for i in range(k):
current_prod *= arr[n-1-i]
max_prod = current_prod
for i in range(k):
current_prod = (current_prod/arr[-(k-i)])*arr[i]
max_prod = max(current_prod, max_prod)
return max_prod
, but have no idea how to return the subsequence (not the product).

Related

how to shrink an array if two consecutive numbers in an array are equal then remove one and increment other

How to shrink an array if two consecutive numbers in an array are equal then remove one and increment other
Example 1:
int a[6]={2,2,3,4,4,4};
// Output: 6
Example 2:
int b[7]={1,2,2,2,4,2,4};
// Output: {1,3,2,4,2,4}
lst = [2,2,3,4,4,4]
def shrink(lst):
idx = 0
while len(lst) > idx+1:
a, b = lst.pop(idx), lst.pop(idx)
if a == b:
lst.insert(idx, a+1)
idx = 0
else:
lst.insert(idx, b)
lst.insert(idx, a)
idx += 1
shrink(lst)
print(lst)
Prints:
[6]
For [5, 5, 5, 1] prints [6, 5, 1]
This can be done in near-linear time like so:
a = [2, 2, 3, 4, 4, 4]
b = [1, 2, 2, 2, 4, 2, 4]
c = [5, 5, 5, 1]
def shrink_array(a):
res = []
for i in range(1, len(a)+1):
if i < len(a) and a[i] == a[i-1]: # if equal to previous
a[i] += 1 # increment and move on
else:
if len(res) > 0 and res[-1] == a[i-1]: # if equal to last in res
res[-1] += 1 # increment last in res
else:
res.append(a[i-1]) # add to res
while len(res) > 1 and res[-1] == res[-2]: # shrink possible duplicates
res[-2] += 1
del res[-1]
return(res)
for arr in [a, b, c]:
print(shrink_array(arr))
Output:
[6]
[1, 3, 2, 4, 2, 4]
[6, 5, 1]

Proper way to make array from integer division in Ruby

tl;dr I want to make an array from the division by 5 results:
20 => [5,5,5,5]
16 => [5,5,5,1]
7 => [5,2]
My current implementation is straightforward yet too large. How can I make it simpler and shorter?
max_count = 5
total_count = input_value
count_array = []
div = total_count / max_count
mod = total_count % max_count
div.times { count_array << max_count }
count_array << mod unless mod == 0
You don't need total_count.
div.times { count_array << max_count } is [max_count] * count_array
Using splat, we can simplify it further
max_count = 5
[*[max_count] * (input_value / max_count), input_value % max_count] - [0]
Alternatively, using divmod
max_count = 5
n, mod = input_value.divmod(max_count)
[*[max_count] * n, mod] - [0]
Last line can also be written as:
(Array.new(n) { max_count } << mod) - [0]
or as Stefan suggested in the comment, using Numeric#nonzero?:
Array.new(n, max_count).push(*mod.nonzero?)
One option more:
d = 5
n = 24
Array.new(n/d){d}.tap{ |a| a << n%d if (n%d).nonzero? }
#=> [5, 5, 5, 5, 4]
You can try this as well.
max=5
num=48
q, r=num.divmod(max) # => [9, 3]
Array.new.fill(max, 0, q).push(r.nonzero?).compact
# => [5, 5, 5, 5, 5, 5, 5, 5, 5, 3]
What about this?
[20].tap{|a| a.push(5, a.pop - 5) while a.last > 5} # => [5, 5, 5, 5]
[16].tap{|a| a.push(5, a.pop - 5) while a.last > 5} # => [5, 5, 5, 1]
[7] .tap{|a| a.push(5, a.pop - 5) while a.last > 5} # => [5, 2]

Removing elements from an array

Problem:
I have two arrays A and B:
A = [0, 1, 2, 3]; %A will always be from 0 to N where N in this case is 3.
B = [0, 1, 3, 1, 9, 4, 6, 2, 5, 9, 10, 11, 3, 8, 1, 5, 9, 10];
weights_B = [3, 4, 5, 6];
I want to compare the first element of A to the first 3 elements of B and the second element of A to the next 4 elements of B. If the elements of A are equal I remove it from B. So in example:
if (A(1) == B(1:3))
remove A(1) from B
Similarly,
I want to compare A(2) to the next 4 elements of B i.e. to B(4:7):
if (A(2) == B(4:7))
remove A(2) from B
I want to compare A(3) to the next 5 elements of B i.e. to B(8:12)
if (A(3) == B(8:12))
remove A(3) from B
I want to compare A(4) to the next 6 elements of B i.e. to B(13:18)
if (A(4) == B(13:18))
remove A(4) from B
Note: The array weights_B determines the number of elements in B that should be respectively compared to A(1), A(2), .. , A(4)
So in the end B should have the following elements:
B = [1, 3, 9, 4, 6, 5, 9, 10, 11, 8, 1, 5, 9, 10];
Needed Solution:
Is there any way I can do this without having to hard-code the indices?
Here's a way without hard-coding:
Bw = mat2cell(B, 1, weights_B); % split into chunks
result = cell(size(Bw)); % initiallize result
for k = 1: numel(A)
result{k} = Bw{k}(Bw{k}~=A(k)); % fill each chunk of the result
end
result = [result{:}]; % concatenate into a row vector
For the sake of diversity, here's a way to do this using splitapply:
function out = q50982235
A = 0:3;
B = [0, 1, 3, 1, 9, 4, 6, 2, 5, 9, 10, 11, 3, 8, 1, 5, 9, 10];
weights_B = [3, 4, 5, 6];
a_ind = 0; % acts as a "global" variable for the inner function
G = repelem( 1:numel(weights_B), weights_B ); % this creates a vector of groups
out = cell2mat( splitapply(#movdif, B, G) );
function out = movdif(B)
a_ind = a_ind + 1;
out = {B(B ~= A(a_ind))};
end
end
The above works because the order of processed groups is predictable.
This solution requires R2015b.
Try this
A = [0, 1, 2, 3];
B = [0, 1, 3, 1, 9, 4, 6, 2, 5, 9, 10, 11, 3, 8, 1, 5, 9, 10];
weights_B = A + A(end);
border_0 = zeros(size(A));
border_1 = zeros(size(A));
border_0(1) = 1;
border_1(end) = length(B);
for i= 2:length(A)
border_0(i) = border_0(i-1) + weights_B(i-1);
border_1(i-1) = border_0(i)-1;
end
C = [];
for i= 1:length(border_0)
shift = 0;
if (i > 1)
shift = border_1(i-1);
end
C = [C B( find(B(border_0(i):border_1(i))~=A(i)) + shift )]
end
A = [0, 1];
B = [0, 1, 3, 1, 4, 5, 6];
% Split B into cells
C{1} = B(1:3) ; % this can be coded if more splits are required
C{2} = B(4:end) ;
% removing the lements
for i = 1:2
C{i}(C{i}==A(i))=[] ; % remove the elements in C{i} present in A(i)
end
cell2mat(C)
Since you want to compare the elements of A with first 3 and then 4 elements of B respectively, you would need to involve indexes.
You could simply use loop for it.
for(int i=0;i<B.length;i++){
if((A[0]==B[i])&&i<3){
B[i]=B[i+1];
}
else if((A[0]==B[i])&&i>3){}
B[i]=B[i+1];
}
Then adjust the updated size of array B.

Find all integers between m and n whose sum of squared divisors is itself a square

Problem Question
Divisors of 42 are : 1, 2, 3, 6, 7, 14, 21, 42. These divisors squared are: 1, 4, 9, 36, 49, 196, 441, 1764. The sum of the squared divisors is 2500 which is 50 * 50, a square!
Given two integers m, n (1 <= m <= n) we want to find all integers between m and n whose sum of squared divisors is itself a square. 42 is such a number.
The result will be an array of arrays, each subarray having two elements, first the number whose squared divisors is a square and then the sum of the squared divisors.
Code below
How can I make this specific program run faster? My current code times out after n > 9999.
#returns the divisors of each number in an array of arrays
r = (m..n).to_a.map { |z| (1..z).select { |x| z % x == 0} }
#this finds all integers between m and n whose sum of squared divisors is itself a square
squarenumbers = r.map { |x| x.map { |c| c**2 }.inject(:+) }.select { |x| Math.sqrt(x) % 1 == 0 }
#returns an array of booleans.
booleans = r.map { |x| x.map { |c| c**2 }.inject(:+) }.map { |x| Math.sqrt(x) % 1 == 0 }
#returns the index of each of the true values in booleans as an array
indexer = booleans.map.with_index{|x, i| i if x == true }.compact
#returns the numbers whose squared divisors is a square in an array
unsqr = indexer.map { |x| (m..n).to_a[x] }
#merges the two arrays together, element for element and creates an array of arrays
unsqr.zip(squarenumbers)
# for m = 1 and n = 1000 the result would be
# [[1, 1], [42, 2500], [246, 84100], [287, 84100], [728, 722500]]
Brute-force calculatioins of factors
You begin by calculating:
m, n = 40, 42
r = (m..n).to_a.map { |z| (1..z).select { |x| z % x == 0} }
#=> [[1, 2, 4, 5, 8, 10, 20, 40], [1, 41], [1, 2, 3, 6, 7, 14, 21, 42]]
That's OK, but you don't need .to_a:
r = (m..n).map { |z| (1..z).select { |x| z % x == 0} }
#=> [[1, 2, 4, 5, 8, 10, 20, 40], [1, 41], [1, 2, 3, 6, 7, 14, 21, 42]]
This avoids an extra step, which is the creation of the temporary array1,2:
(m..n).to_a #=> [40, 41, 42]
Structure of a solution
Let's work backwards to come up with our code. First, concentrate on determining, for any given number q, if the sum of squares of the factors of q is itself a perfect square. Suppose we construct a method magic_number? which takes q as its only argument and returns true if q satisfies the required property and false otherwise. Then we will compute:
(m..n).select { |q| magic_number?(q) }
to return an array of all numbers between m and n that satisfy the property. magic_number? can be written like this:
def magic_number?(q)
return true if q == 1
s = sum_of_squared_factors(q)
s == Math.sqrt(s).round**2
end
Calculating sum of squared factors
So now we are left with writing the method sum_of_squared_factors. We can use your code to obtain the factors:
def factors(q)
(1..q).select { |x| q % x == 0 }
end
factors(40) #=> [1, 2, 4, 5, 8, 10, 20, 40]
factors(41) #=> [1, 41]
factors(42) #=> [1, 2, 3, 6, 7, 14, 21, 42]
and then write:
def sum_of_squared_factors(q)
factors(q).reduce(0) { |t,i| t + i*i }
end
sum_of_squared_factors(40) #=> 2210
sum_of_squared_factors(41) #=> 1682
sum_of_squared_factors(42) #=> 2500
Speeding the calculation of factors
There's something more we can do to speed up the calculation of factors. If f is a factor of n, f and n/f, are both factors of n. (For example, since 3 is a factor of 42, so is 42/3 #=> 14). We therefore need only obtain the smaller of each pair.
There is one exception to this rule. If n is a perfect square and f == n**0.5, then f = n/f, so we only include f among the factors of n (not n/f as well).
If turns out that if f is the smaller of the pair, f <=(n**0.5).round3. We therefore need only check to see which of the numbers (1..(n**0.5).round) are factors and include their complements (unless n is a perfect square, in which case we do not double-count (n**0.5).round):
q = 42
arr = (1..Math.sqrt(q).round).select { |x| q % x == 0 }
#=> [1, 2, 3, 6]
arr = arr.flat_map { |n| [n, q/n] }
#=> [1, 42, 2, 21, 3, 14, 6, 7]
arr.pop if a[-2] == a[-1]
arr
#=> [1, 42, 2, 21, 3, 14, 6, 7]
q = 36
arr = (1..Math.sqrt(q).round).select { |x| q % x == 0 }
#=> [1, 2, 3, 4, 6]
arr = arr.flat_map { |n| [n, q/n] }
#=> [1, 36, 2, 18, 3, 12, 4, 9, 6, 6]
arr.pop if a[-2] == a[-1]
#=> 6
arr
#=> [1, 36, 2, 18, 3, 12, 4, 9, 6]
so we can write:
def factors(q)
arr = (1..Math.sqrt(q)).select { |x| q % x == 0 }
arr = arr.flat_map { |n| [n, q/n] }
arr.pop if arr[-2] == arr[-1]
arr
end
Substituting out arr ("chaining" expressions), we obtain a typical Ruby expression:
def factors(q)
(1..Math.sqrt(q)).select { |x| q % x == 0 }.
flat_map { |n| [n, q/n] }.
tap { |a| a.pop if a[-2] == a[-1] }
end
factors(42)
#=> [1, 42, 2, 21, 3, 14, 6, 7]
factors(36)
#=> [1, 36, 2, 18, 3, 12, 4, 9, 6]
See Enumerable#flat_map and Object#tap. (There's no need for this array to be sorted. In applications where it needs to be sorted, just tack .sort onto the end of flat_maps block.)
Wrapping up
In sum, we are left with the following:
def magic_number?(q)
return true if q == 1
s = sum_of_squared_factors(q)
s == Math.sqrt(s).round**2
end
def sum_of_squared_factors(q)
factors(q).reduce(0) { |t,i| t + i*i }
end
def factors(q)
(1..Math.sqrt(q)).select { |x| q % x == 0 }.
flat_map { |n| [n, q/n] }.
tap { |a| a.pop if a[-2] == a[-1] }
end
m, n = 1, 1000
(m..n).select { |q| magic_number?(q) }
#=> `[1, 42, 246, 287, 728]
This calculation was completed in a blink of an eye.
Compute primes to further speed calculation of factors
Lastly, let me describe an even faster way to compute the factors of a number, using the method Prime::prime_division. That method decomposes any number into its prime components. Consider, for example, n = 360.
require 'prime'
Prime.prime_division(360)
#=> [[2, 3], [3, 2], [5, 1]]
This tells us that:
360 == 2**3 * 3**2 * 5**1
#=> true
It also tells us that every factor of 360 is the product of between 0 and 3 2's, multiplied by between 0 and 2 3's, multiplied by 0 or 1 5's. Therefore:
def factors(n)
Prime.prime_division(n).reduce([1]) do |a,(prime,pow)|
a.product((0..pow).map { |po| prime**po }).map { |x,y| x*y }
end
end
a = factors(360).sort
#=> [ 1, 2, 3, 4, 5, 6, 8, 9, 10, 12, 15, 18,
# 20, 24, 30, 36, 40, 45, 60, 72, 90, 120, 180, 360]
We can check that:
a == (1..360).select { |n| (360 % n).zero? }
#=> true
One other check:
factors(40).sort
#=> [1, 2, 4, 5, 8, 10, 20, 40]
1. You could instead write that [*m..n] #=> [40, 41, 42].
2. Why is it not necessary to convert the range to an array? Enumerable#map, being an instance method of the module Enumerable, is available for use by every class that includes Enumerable. Array is one, but (m..n).class #=> Range is another. (See the second paragraph at Range).
3. Suppose f is smaller than n/f and f > n**0.5, then n/f < n/(n**0.5) = n**0.5 < f, a contradiction.
I don't know Ruby but the problem lies with the algorithm used in finding the divisors of a number (which is not specific to the language used, i.e. Ruby in this case).
r = (m..n).to_a.map { |z| (1..z).select { |x| z % x == 0} }
To find the divisors of an integer n you are dividing n by all positive integers unto n - 1 which means the loop runs n - 1 times. However, it is enough to divide upto sort(n) to calculate the divisors. In pseudocode this looks like below:
for i = 1 to i <= sqrt(n)
r = n % i
if r == 0 then
i is a divisor
if n / i != i then
n / i is another divisor
For example:
sqrt_42 = 6.48074069840786
i = 1 => 1 and 42 are two divisors
i = 2 => 2 and 21
i = 3 => 3 and 14
i = 4 => no divisor
i = 5 => no divisor
i = 6 => 6 and 7
And thats all.
This will improve the performance a lot since now the loop runs only sort(n) times instead of n - 1 times which is a big difference for large n.

Filling missing data in a data set with constant values

I have a data set like the following:
x= [1, 4, 10]
y= [10, 20, 30]
(x and y are value pairs, i.e. (1,10), (4,20), (10,30))
I would like to fill the x values gaps and having constant values for y until the next known value pair comes.This should be done between each value pair, i.e. between (1,10) and (4,20) and then again between (4,20) and (10,30).
Input:
x=[1, 4, 10];
y=[10, 20, 30];
Output:
xi= [1, 2, 3, 4, 5, 6, 7, 8, 9, 10];
yi= [10,10, 10, 20, 20, 20, 20, 20, 20, 30];
How can Matlab solve this for me?
Assuming ascending order of elements in x, this could be one approach based on diff & cumsum -
%// Sample inputs
x=[1, 4, 10]
y=[-2, 5, -3]
xi = min(x):max(x)
yi = zeros(1,numel(xi))
yi(x) = diff([0 y])
yi = cumsum(yi)
Sample run -
x =
1 4 10
y =
-2 5 -3
xi =
1 2 3 4 5 6 7 8 9 10
yi =
-2 -2 -2 5 5 5 5 5 5 -3
Customary bsxfun solution to get yi -
lens = [diff(x) 1];
yi = nonzeros(bsxfun(#times,bsxfun(#ge,lens,[1:max(lens)]'),y)).'
Assuming that x always starts with a 1 and finishes with the final length of xi, this will work:
xi=1:x(end)
yi=y(arrayfun(#(xi)find(x<=xi,1,'last'),xi))

Resources