We Keep Coding

Remove Triplet from array from last

Given an integer array of nums, remove the triplet from last.
The relative order of elements should be kept the same.
Example 1
input : nums = [2,4,2,2,7,5,6,7,8,6,6,2,6,7,6]
output : nums = [2,4,5,6,8,6]
Example 2 input : nums = [2,2,3,2,3,2]
output : nums = [2,3,3]
I have this in java
int[] nums = {2,4,2,2,7,5,6,7,8,6,6,2,6,7,6};
int[] ans = new int[6];
int count=1;
for(int i=0;i<nums.length;i++){
for(int j=0;j<nums.length;j++){
if(arr[i] == arr[j]){
if(count < 3){
count++;
ans[i] = nums[i];
}
}
}
}

I couldn't let this one go, so I did code up an example in c#. This is one approach. There are probably many ways to do this better:
private static int[] RemoveLastTriplet(int[] input)
{
var toRemove = new bool[input.Length];
var counts = new Dictionary<int, int>();
//Count how many times each input value is found.
foreach(var value in input)
{
int count;
if (counts.TryGetValue(value, out count))
{
counts[value] = count + 1;
}
else
{
counts[value] = 1;
}
}
foreach(var kvp in counts)
{
//Determine how many triplets we have for this value
var tripletCount = kvp.Value / 3;
//Keep track of where we're starting
var currentIndex = input.Length - 1;
//Remove each triplet
for(var tripletIndex = 0; tripletIndex < tripletCount; tripletIndex++)
{
//counts the number of elements in this triplet
var thisTripletCount = 0;
//Mark each member of the triplet for deletion.
for(var inputIndex = currentIndex; thisTripletCount < 3; inputIndex--)
{
if (input[inputIndex] == kvp.Key)
{
//Mark this index for removal
toRemove[inputIndex] = true;
thisTripletCount++;
}
//Keep track of where we are in the overall input array
currentIndex--;
}
}
}
//We could be more clever here and keep track of how many
// items we'll have in the output list and just create an array.
var output = new List<int>();
for(int index = 0; index < input.Length; index++)
{
if (!toRemove[index])
{
output.Add(input[index]);
}
}
return output.ToArray();
}

This is my n^2 solution for this problem. I hope it will help you. Just find the count of element first and then find the total number of elements to be removed. Let say it is 't' after that, just replace the t number of elements from the end with -1.
#include <iostream>
#include <unordered_set>
using namespace std;
int main() {
int n;
cout << "Enter the length of array : ";
cin >> n;
int arr[n];
cout << "Enter items : ";
for (int i = 0; i < n; ++i)
cin >> arr[i];
// set for fast lookups
unordered_set<int> uset;
for (int i = n - 1; i > -1; --i) {
int e = arr[i];
if (uset.find(e) != uset.end())
continue;
// count the element frequiency
int c = 0;
for (int j = 0; j < n; ++j) {
if (arr[j] == e)
c++;
}
// number of elements would we remove from end
c = c - c % 3;
for (int j = n - 1; j > -1; --j) {
if (c == 0)
break;
if (arr[j] == e) {
arr[j] = -1; // put -1 at removed element location
c--;
}
}
}
// now just print only elemt those value is not equals to zero
for (int i = 0; i < n; ++i) {
if (arr[i] != (-1))
cout << arr[i] << " ";
}
cout << endl;
return 0;
}

arr = input().split()[::-1]
# print(arr)
i = 0
def count_same(num, ind):
count = 0
indexes = []
for index in range(ind, len(arr)):
if arr[index] == num:
count += 1
indexes.append(index)
if count == 3:
break
if count == 3:
for index in indexes[::-1]:
arr.pop(index)
return True
return False
while i < len(arr):
if not count_same(arr[i], i):
i += 1
# print(arr)
print(list(map(int, arr[::-1])))

#include<bits/stdc++.h>
using namespace std;
int main(){
unordered_map<int, int> mp;
vector<int> nums = {2,2,3,2,3,2};
for(auto it:nums){
mp[it]++;
}
for (int i = 0; i < nums.size();i++){
if(mp[nums[i]]%3==0){
continue;
}
else{
cout << nums[i] << " ";
mp[nums[i]] -= mp[nums[i]] / 3;
}
}
}

These python code is working you can use the below exapmles
Example 1:
Input: nums = [2,4,2,2,7,5,6,7,8,6,6,2,6,7,6]
Output: nums = [2,4,5,6,8,6]
Example 2:
Input: nums[2,2,3,2,3,2]
Output: nums[2,3,3]
from collections import defaultdict
nums = list(map(int, input("Enter the elements of the array separated by commas: ").split(',')))
mp = defaultdict(int)
for i in nums:
mp[i] += 1
result = []
for i in nums:
if mp[i] % 3 != 0:
result.append(i)
mp[i] -= mp[i] // 3
print(result)

Sorting only evens in an array while keeping the odds in place WITHOUT creating a new array [closed]

Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed last year.
Improve this question
Is there any way to do this?
Doing it with a separate array to hold and sort all of the evens is pretty trivial:
place all evens in a separate array
sort the array containing the evens
iterate through original array, replacing the even elements
But... I can't seem to find a way to do this without a separate array to hold the evens. Any tips?

Sure. I'm confident many of the standard in-place search algorithms can be modified, but insertion sort is particularly easy. The usual sort:
void sort(int *a, int size) {
for (int s = 1; s < size; ++s) {
int t = a[s], i;
for (i = s; i > 0 && t < a[i - 1]; --i) a[i] = a[i - 1];
a[i] = t;
}
}
So here we do the same thing except with extra inserted loops to ignore odd entries. This is a quick hack, very possibly with bugs remaining:
void sort_evens(int *a, int size) {
int i, s;
for (s = 0; s < size && a[s] % 2 == 1; ++s) /* skip */;
if (s == size) return;
while (1) {
for (++s; s < size && a[s] % 2 == 1; ++s) /* skip */;
if (s == size) return;
int i_last = s, t = a[s];
for (i = s - 1; i >= 0; --i) {
for ( ; i >= 0 && a[i] % 2 == 1; --i) /* skip */;
if (i < 0 || t >= a[i]) break;
a[i_last] = a[i];
i_last = i;
}
a[i_last] = t;
}
}
On this data:
int a[] = {7,3,10,9,0,1,4,3,2,1,6,8,3,5,};
it produces
7 3 0 9 2 1 4 3 6 1 8 10 3 5
More fun would be heapsort.

You would use any sorting algorithm and ignore the Odd ones.
Just swap values in-place without creating a new array.
for (int i = 0; i < input.length; i++)
{
var cur = input[i];
// ignore odd
if (cur % 2 != 0)
continue;
var lowest = cur;
var lowestIx = 0;
// find the lowest value
for (int j = i + 1; j < input.length; j++)
{
var next = input[j];
// ignore odd
if (next % 2 != 0)
continue;
if (next < lowest)
{
lowestIx = j;
lowest = next;
}
}
// swap with lowest
if (lowest != cur)
{
var tmp = input[lowestIx];
input[lowestIx] = input[i];
input[i] = tmp;
}
}

How to find the longest substring with equal amount of characters efficiently

I have a string that consists of characters A,B,C and D and I am trying to calculate the length of the longest substring that has an equal amount of each one of these characters in any order.
For example ABCDB would return 4, ABCC 0 and ADDBCCBA 8.
My code currently:
public int longestSubstring(String word) {
HashMap<Integer, String> map = new HashMap<Integer, String>();
for (int i = 0; i<word.length()-3; i++) {
map.put(i, word.substring(i, i+4));
}
StringBuilder sb;
int longest = 0;
for (int i = 0; i<map.size(); i++) {
sb = new StringBuilder();
sb.append(map.get(i));
int a = 4;
while (i<map.size()-a) {
sb.append(map.get(i+a));
a+= 4;
}
String substring = sb.toString();
if (equalAmountOfCharacters(substring)) {
int length = substring.length();
if (length > longest)
longest = length;
}
}
return longest;
}
This currently works pretty well if the string length is 10^4 but I'm trying to make it 10^5. Any tips or suggestions would be appreciated.

Let's assume that cnt(c, i) is the number of occurrences of the character c in the prefix of length i.
A substring (low, high] has an equal amount of two characters a and b iff cnt(a, high) - cnt(a, low) = cnt(b, high) - cnt(b, low), or, put it another way, cnt(b, high) - cnt(a, high) = cnt(b, low) - cnt(a, low). Thus, each position is described by a value of cnt(b, i) - cnt(a, i). Now we can generalize it for more that two characters: each position is described by a tuple (cnt(a_2, i) - cnt(a_1, i), ..., cnt(a_k, i) - cnt(a_1, i)), where a_1 ... a_k is the alphabet.
We can iterate over the given string and maintain the current tuple. At each step, we should update the answer by checking the value of i - first_occurrence(current_tuple), where first_occurrence is a hash table that stores the first occurrence of each tuple seen so far. Do not forget to put a tuple of zeros to the hash map before iteration(it corresponds to an empty prefix).

If there were only A's and B's, then you could do something like this.
def longest_balanced(word):
length = 0
cumulative_difference = 0
first_index = {0: -1}
for index, letter in enumerate(word):
if letter == 'A':
cumulative_difference += 1
elif letter == 'B':
cumulative_difference -= 1
else:
raise ValueError(letter)
if cumulative_difference in first_index:
length = max(length, index - first_index[cumulative_difference])
else:
first_index[cumulative_difference] = index
return length
Life is more complicated with all four letters, but the idea is much the same. Instead of keeping just one cumulative difference, for A's versus B's, we keep three, for A's versus B's, A's versus C's, and A's versus D's.

Well, first of all abstain from constructing any strings.
If you don't produce any (or nearly no) garbage, there's no need to collect it, which is a major plus.
Next, use a different data-structure:
I suggest 4 byte-arrays, storing the count of their respective symbol in the 4-span starting at the corresponding string-index.
That should speed it up considerably.

You can count the occurrences of the characters in word. Then, a possible solution could be:
If min is the minimum number of occurrences of any character in word, then min is also the maximum possible number of occurrences of each character in the substring we are looking for. In the code below, min is maxCount.
We iterate over decreasing values of maxCount. At every step, the string we are searching for will have length maxCount * alphabetSize. We can view this as the size of a sliding window we can slide over word.
We slide the window over word, counting the occurrences of the characters in the window. If the window is the substring we are searching for, we return the result. Otherwise, we keep searching.
[FIXED] The code:
private static final int ALPHABET_SIZE = 4;
public int longestSubstring(String word) {
// count
int[] count = new int[ALPHABET_SIZE];
for (int i = 0; i < word.length(); i++) {
char c = word.charAt(i);
count[c - 'A']++;
}
int maxCount = word.length();
for (int i = 0; i < count.length; i++) {
int cnt = count[i];
if (cnt < maxCount) {
maxCount = cnt;
}
}
// iterate over maxCount until found
boolean found = false;
while (maxCount > 0 && !found) {
int substringLength = maxCount * ALPHABET_SIZE;
found = findSubstring(substringLength, word, maxCount);
if (!found) {
maxCount--;
}
}
return found ? maxCount * ALPHABET_SIZE : 0;
}
private boolean findSubstring(int length, String word, int maxCount) {
int startIndex = 0;
boolean found = false;
while (startIndex + length <= word.length()) {
int[] count = new int[ALPHABET_SIZE];
for (int i = startIndex; i < startIndex + length; i++) {
char c = word.charAt(i);
int cnt = ++count[c - 'A'];
if (cnt > maxCount) {
break;
}
}
if (equalValues(count, maxCount)) {
found = true;
break;
} else {
startIndex++;
}
}
return found;
}
// Returns true if all values in c are equal to value
private boolean equalValues(int[] count, int value) {
boolean result = true;
for (int i : count) {
if (i != value) {
result = false;
break;
}
}
return result;
}
[MERGED] This is Hollis Waite's solution using cumulative counts, but taking my observations at points 1. and 2. into consideration. This may improve performance for some inputs:
private static final int ALPHABET_SIZE = 4;
public int longestSubstring(String word) {
// count
int[][] cumulativeCount = new int[ALPHABET_SIZE][];
for (int i = 0; i < ALPHABET_SIZE; i++) {
cumulativeCount[i] = new int[word.length() + 1];
}
int[] count = new int[ALPHABET_SIZE];
for (int i = 0; i < word.length(); i++) {
char c = word.charAt(i);
count[c - 'A']++;
for (int j = 0; j < ALPHABET_SIZE; j++) {
cumulativeCount[j][i + 1] = count[j];
}
}
int maxCount = word.length();
for (int i = 0; i < count.length; i++) {
int cnt = count[i];
if (cnt < maxCount) {
maxCount = cnt;
}
}
// iterate over maxCount until found
boolean found = false;
while (maxCount > 0 && !found) {
int substringLength = maxCount * ALPHABET_SIZE;
found = findSubstring(substringLength, word, maxCount, cumulativeCount);
if (!found) {
maxCount--;
}
}
return found ? maxCount * ALPHABET_SIZE : 0;
}
private boolean findSubstring(int length, String word, int maxCount, int[][] cumulativeCount) {
int startIndex = 0;
int endIndex = (startIndex + length) - 1;
boolean found = true;
while (endIndex < word.length()) {
for (int i = 0; i < ALPHABET_SIZE; i++) {
if (cumulativeCount[i][endIndex] - cumulativeCount[i][startIndex] != maxCount) {
found = false;
break;
}
}
if (found) {
break;
} else {
startIndex++;
endIndex++;
}
}
return found;
}

You'll probably want to cache cumulative counts of characters for each index of String -- that's where the real bottleneck is. Haven't thoroughly tested but something like the below should work.
public class Test {
static final int LEN = 4;
static class RandomCharSequence implements CharSequence {
private final Random mRandom = new Random();
private final int mAlphabetLen;
private final int mLen;
private final int mOffset;
RandomCharSequence(int pLen, int pOffset, int pAlphabetLen) {
mAlphabetLen = pAlphabetLen;
mLen = pLen;
mOffset = pOffset;
}
public int length() {return mLen;}
public char charAt(int pIdx) {
mRandom.setSeed(mOffset + pIdx);
return (char) (
'A' +
(mRandom.nextInt() % mAlphabetLen + mAlphabetLen) % mAlphabetLen
);
}
public CharSequence subSequence(int pStart, int pEnd) {
return new RandomCharSequence(pEnd - pStart, pStart, mAlphabetLen);
}
#Override public String toString() {
return (new StringBuilder(this)).toString();
}
}
public static void main(String[] pArgs) {
Stream.of("ABCDB", "ABCC", "ADDBCCBA", "DADDBCCBA").forEach(
pWord -> System.out.println(longestSubstring(pWord))
);
for (int i = 0; ; i++) {
final double len = Math.pow(10, i);
if (len >= Integer.MAX_VALUE) break;
System.out.println("Str len 10^" + i);
for (int alphabetLen = 1; alphabetLen <= LEN; alphabetLen++) {
final Instant start = Instant.now();
final int val = longestSubstring(
new RandomCharSequence((int) len, 0, alphabetLen)
);
System.out.println(
String.format(
" alphabet len %d; result %08d; time %s",
alphabetLen,
val,
formatMillis(ChronoUnit.MILLIS.between(start, Instant.now()))
)
);
}
}
}
static String formatMillis(long millis) {
return String.format(
"%d:%02d:%02d.%03d",
TimeUnit.MILLISECONDS.toHours(millis),
TimeUnit.MILLISECONDS.toMinutes(millis) -
TimeUnit.HOURS.toMinutes(TimeUnit.MILLISECONDS.toHours(millis)),
TimeUnit.MILLISECONDS.toSeconds(millis) -
TimeUnit.MINUTES.toSeconds(TimeUnit.MILLISECONDS.toMinutes(millis)),
TimeUnit.MILLISECONDS.toMillis(millis) -
TimeUnit.SECONDS.toMillis(TimeUnit.MILLISECONDS.toSeconds(millis))
);
}
static int longestSubstring(CharSequence pWord) {
// create array that stores cumulative char counts at each index of string
// idx 0 = char (A-D); idx 1 = offset
final int[][] cumulativeCnts = new int[LEN][];
for (int i = 0; i < LEN; i++) {
cumulativeCnts[i] = new int[pWord.length() + 1];
}
final int[] cumulativeCnt = new int[LEN];
for (int i = 0; i < pWord.length(); i++) {
cumulativeCnt[pWord.charAt(i) - 'A']++;
for (int j = 0; j < LEN; j++) {
cumulativeCnts[j][i + 1] = cumulativeCnt[j];
}
}
final int maxResult = Arrays.stream(cumulativeCnt).min().orElse(0) * LEN;
if (maxResult == 0) return 0;
int result = 0;
for (int initialOffset = 0; initialOffset < LEN; initialOffset++) {
for (
int start = initialOffset;
start < pWord.length() - result;
start += LEN
) {
endLoop:
for (
int end = start + result + LEN;
end <= pWord.length() && end - start <= maxResult;
end += LEN
) {
final int substrLen = end - start;
final int expectedCharCnt = substrLen / LEN;
for (int i = 0; i < LEN; i++) {
if (
cumulativeCnts[i][end] - cumulativeCnts[i][start] !=
expectedCharCnt
) {
continue endLoop;
}
}
if (substrLen > result) result = substrLen;
}
}
}
return result;
}
}

Suppose there are K possible letters in a string of length N. We could track the balance of letters seen with a vector pos of length K that is updated as follows:
If letter 1 is seen, add (K-1, -1, -1, ...)
If letter 2 is seen, add (-1, K-1, -1, ...)
If letter 3 is seen, add (-1, -1, K-1, ...)
Maintain a hash that maps pos to the first string position where pos is reached. Balanced substrings occur whenever hash[pos] already exists and the substring value is s[hash[pos]:pos].
The cost of maintaining the hash is O(log N) so processing the string takes O(N log N). How does this compare with solutions so far? These types of problems tend to have linear solutions but I haven't come across one yet.
Here's some code demonstrating the idea for 3 letters and a run using biased random strings. (Uniform random strings allow for solutions that are around half the string length, which is unwieldy to print).
#!/usr/bin/python
import random
from time import time
alphabet = "abc"
DIM = len(alphabet)
def random_string(n):
# return a random string over choices[] of length n
# distribution of letters is non-uniform to make matches harder to find
choices = "aabbc"
s = ''
for i in range(n):
r = random.randint(0, len(choices) - 1)
s += choices[r]
return s
def validate(s):
# verify frequencies of each letter are the same
f = [0, 0, 0]
a2f = {alphabet[i] : i for i in range(DIM)}
for c in s:
f[a2f[c]] += 1
assert f[0] == f[1] and f[1] == f[2]
def longest_balanced(s):
"""return length of longest substring of s containing equal
populations of each letter in alphabet"""
slen = len(s)
p = [0 for i in range(DIM)]
vec = {alphabet[0] : [2, -1, -1],
alphabet[1] : [-1, 2, -1],
alphabet[2] : [-1, -1, 2]}
x = -1
best = -1
hist = {str([0, 0, 0]) : -1}
for c in s:
x += 1
p = [p[i] + vec[c][i] for i in range(DIM)]
pkey = str(p)
if pkey not in hist:
hist[pkey] = x
else:
span = x - hist[pkey]
assert span % DIM == 0
if span > best:
best = span
cand = s[hist[pkey] + 1: x + 1]
print("best so far %d = [%d,%d]: %s" % (best,
hist[pkey] + 1,
x + 1,
cand))
validate(cand)
return best if best > -1 else 0
def main():
#print longest_balanced( "aaabcabcbbcc" )
t0 = time()
s = random_string(1000000)
print "generate time:", time() - t0
t1 = time()
best = longest_balanced( s )
print "best:", best
print "elapsed:", time() - t1
main()
Sample run on an input of 10^6 letters with an alphabet of 3 letters:
$ ./bal.py
...
best so far 189 = [847894,848083]: aacacbcbabbbcabaabbbaabbbaaaacbcaaaccccbcbcbababaabbccccbbabbacabbbbbcaacacccbbaacbabcbccaabaccabbbbbababbacbaaaacabcbabcbccbabbccaccaabbcabaabccccaacccccbaacaaaccbbcbcabcbcacaabccbacccacca
best: 189
elapsed: 1.43609690666

java codility Max-Counters

I have been trying to solve the below task:
You are given N counters, initially set to 0, and you have two possible operations on them:
increase(X) − counter X is increased by 1,
max_counter − all counters are set to the maximum value of any counter.
A non-empty zero-indexed array A of M integers is given. This array represents consecutive operations:
if A[K] = X, such that 1 ≤ X ≤ N, then operation K is increase(X),
if A[K] = N + 1 then operation K is max_counter.
For example, given integer N = 5 and array A such that:
A[0] = 3
A[1] = 4
A[2] = 4
A[3] = 6
A[4] = 1
A[5] = 4
A[6] = 4
the values of the counters after each consecutive operation will be:
(0, 0, 1, 0, 0)
(0, 0, 1, 1, 0)
(0, 0, 1, 2, 0)
(2, 2, 2, 2, 2)
(3, 2, 2, 2, 2)
(3, 2, 2, 3, 2)
(3, 2, 2, 4, 2)
The goal is to calculate the value of every counter after all operations.
struct Results {
int * C;
int L;
};
Write a function:
struct Results solution(int N, int A[], int M);
that, given an integer N and a non-empty zero-indexed array A consisting of M integers, returns a sequence of integers representing the values of the counters.
The sequence should be returned as:
a structure Results (in C), or
a vector of integers (in C++), or
a record Results (in Pascal), or
an array of integers (in any other programming language).
For example, given:
A[0] = 3
A[1] = 4
A[2] = 4
A[3] = 6
A[4] = 1
A[5] = 4
A[6] = 4
the function should return [3, 2, 2, 4, 2], as explained above.
Assume that:
N and M are integers within the range [1..100,000];
each element of array A is an integer within the range [1..N + 1].
Complexity:
expected worst-case time complexity is O(N+M);
expected worst-case space complexity is O(N), beyond input storage (not counting the storage required for input arguments).
Elements of input arrays can be modified.
Here is my solution:
import java.util.Arrays;
class Solution {
public int[] solution(int N, int[] A) {
final int condition = N + 1;
int currentMax = 0;
int countersArray[] = new int[N];
for (int iii = 0; iii < A.length; iii++) {
int currentValue = A[iii];
if (currentValue == condition) {
Arrays.fill(countersArray, currentMax);
} else {
int position = currentValue - 1;
int localValue = countersArray[position] + 1;
countersArray[position] = localValue;
if (localValue > currentMax) {
currentMax = localValue;
}
}
}
return countersArray;
}
}
Here is the code valuation:
https://codility.com/demo/results/demo6AKE5C-EJQ/
Can you give me a hint what is wrong with this solution?

The problem comes with this piece of code:
for (int iii = 0; iii < A.length; iii++) {
...
if (currentValue == condition) {
Arrays.fill(countersArray, currentMax);
}
...
}
Imagine that every element of the array A was initialized with the value N+1. Since the function call Arrays.fill(countersArray, currentMax) has a time complexity of O(N) then overall your algorithm will have a time complexity O(M * N). A way to fix this, I think, instead of explicitly updating the whole array A when the max_counter operation is called you may keep the value of last update as a variable. When first operation (incrementation) is called you just see if the value you try to increment is larger than the last_update. If it is you just update the value with 1 otherwise you initialize it to last_update + 1. When the second operation is called you just update last_update to current_max. And finally, when you are finished and try to return the final values you again compare each value to last_update. If it is greater you just keep the value otherwise you return last_update
class Solution {
public int[] solution(int N, int[] A) {
final int condition = N + 1;
int currentMax = 0;
int lastUpdate = 0;
int countersArray[] = new int[N];
for (int iii = 0; iii < A.length; iii++) {
int currentValue = A[iii];
if (currentValue == condition) {
lastUpdate = currentMax
} else {
int position = currentValue - 1;
if (countersArray[position] < lastUpdate)
countersArray[position] = lastUpdate + 1;
else
countersArray[position]++;
if (countersArray[position] > currentMax) {
currentMax = countersArray[position];
}
}
}
for (int iii = 0; iii < N; iii++) {
if (countersArray[iii] < lastUpdate)
countersArray[iii] = lastUpdate;
}
return countersArray;
}
}

The problem is that when you get lots of max_counter operations you get lots of calls to Arrays.fill which makes your solution slow.
You should keep a currentMax and a currentMin:
When you get a max_counter you just set currentMin = currentMax.
If you get another value, let's call it i:
If the value at position i - 1 is smaller or equal to currentMin you set it to currentMin + 1.
Otherwise you increment it.
At the end just go through the counters array again and set everything less than currentMin to currentMin.

Another solution that I have developed and might be worth considering: http://codility.com/demo/results/demoM658NU-DYR/

This is the 100% solution of this question.
// you can also use imports, for example:
// import java.math.*;
class Solution {
public int[] solution(int N, int[] A) {
int counter[] = new int[N];
int n = A.length;
int max=-1,current_min=0;
for(int i=0;i<n;i++){
if(A[i]>=1 && A[i]<= N){
if(counter[A[i] - 1] < current_min) counter[A[i] - 1] = current_min;
counter[A[i] - 1] = counter[A[i] - 1] + 1;
if(counter[A[i] - 1] > max) max = counter[A[i] - 1];
}
else if(A[i] == N+1){
current_min = max;
}
}
for(int i=0;i<N;i++){
if(counter[i] < current_min) counter[i] = current_min;
}
return counter;
}
}

I'm adding another Java 100 solution with some test cases it they're helpful.
// https://codility.com/demo/results/demoD8J6M5-K3T/ 77
// https://codility.com/demo/results/demoSEJHZS-ZPR/ 100
public class MaxCounters {
// Some testcases
// (1,[1,2,3]) = [1]
// (1,[1]) = [1]
// (1,[5]) = [0]
// (1,[1,1,1,2,3]) = 3
// (2,[1,1,1,2,3,1]) = [4,3]
// (5, [3, 4, 4, 5, 1, 4, 4]) = (1, 0, 1, 4, 1)
public int[] solution(int N, int[] A) {
int length = A.length, maxOfCounter = 0, lastUpdate = 0;
int applyMax = N + 1;
int result[] = new int[N];
for (int i = 0; i < length; ++i ) {
if(A[i] == applyMax){
lastUpdate = maxOfCounter;
} else if (A[i] <= N) {
int position = A[i]-1;
result[position] = result[position] > lastUpdate
? result[position] + 1 : lastUpdate + 1;
// updating the max for future use
if(maxOfCounter <= result[position]) {
maxOfCounter = result[position];
}
}
}
// updating all the values that are less than the lastUpdate to the max value
for (int i = 0; i < N; ++i) {
if(result[i] < lastUpdate) {
result[i] = lastUpdate;
}
}
return result;
}
}

My java solution with a detailed explanation 100% Correctness, 100% Performance :
Time Complexity O(N+M)
public static int[] solution(int N, int[] A) {
int[] counters = new int[N];
//The Max value between all counters at a given time
int max = 0;
//The base Max that all counter should have after the "max counter" operation happens
int baseMax = 0;
for (int i = 0; i < A.length; i++) {
//max counter Operation ==> updating the baseMax
if (A[i] > N) {
// Set The Base Max that all counters should have
baseMax = max;
}
//Verify if the value is bigger than the last baseMax because at any time a "max counter" operation can happen and the counter should have the max value
if (A[i] <= N && counters[A[i] - 1] < baseMax) {
counters[A[i] - 1] = baseMax;
}
//increase(X) Operation => increase the counter value
if (A[i] <= N) {
counters[A[i] - 1] = counters[A[i] - 1] + 1;
//Update the max
max = Math.max(counters[A[i] - 1], max);
}
}
//Set The remaining values to the baseMax as not all counters are guaranteed to be affected by an increase(X) operation in "counters[A[i] - 1] = baseMax;"
for (int j = 0; j < N; j++) {
if (counters[j] < baseMax)
counters[j] = baseMax;
}
return counters;
}

Here is my C++ solution which got 100 on codility. The concept is same as explained above.
int maxx=0;
int lastvalue=0;
void set(vector<int>& A, int N,int X)
{
for ( int i=0;i<N;i++)
if(A[i]<lastvalue)
A[i]=lastvalue;
}
vector<int> solution(int N, vector<int> &A) {
// write your code in C++11
vector<int> B(N,0);
for(unsigned int i=0;i<A.size();i++)
{
if(A[i]==N+1)
lastvalue=maxx;
else
{ if(B[A[i]-1]<lastvalue)
B[A[i]-1]=lastvalue+1;
else
B[A[i]-1]++;
if(B[A[i]-1]>maxx)
maxx=B[A[i]-1];
}
}
set(B,N,maxx);
return B;
}

vector<int> solution(int N, vector<int> &A)
{
std::vector<int> counters(N);
auto max = 0;
auto current = 0;
for (auto& counter : A)
{
if (counter >= 1 && counter <= N)
{
if (counters[counter-1] < max)
counters[counter - 1] = max;
counters[counter - 1] += 1;
if (counters[counter - 1] > current)
current = counters[counter - 1];
}
else if (counter > N)
max = current;
}
for (auto&& counter : counters)
if (counter < max)
counter = max;
return counters;
}

Arrays.fill() invocation inside array interation makes the program O(N^2)
Here is a possible solution which has O(M+N) runtime.
The idea is -
For the second operation, keep track of max value that is achieved through increment, this is our base value till the current iteration, no values can't be less than this.
For the first operation, resetting the value to base value if needed before the increment.
public static int[] solution(int N, int[] A) {
int counters[] = new int[N];
int base = 0;
int cMax = 0;
for (int a : A) {
if (a > counters.length) {
base = cMax;
} else {
if (counters[a - 1] < base) {
counters[a - 1] = base;
}
counters[a - 1]++;
cMax = Math.max(cMax, counters[a - 1]);
}
}
for (int i = 0; i < counters.length; i++) {
if (counters[i] < base) {
counters[i] = base;
}
}
return counters;
}

vector<int> solution(int N, vector<int> &A)
{
std::vector<int> counter(N, 0);
int max = 0;
int floor = 0;
for(std::vector<int>::iterator i = A.begin();i != A.end(); i++)
{
int index = *i-1;
if(*i<=N && *i >= 1)
{
if(counter[index] < floor)
counter[index] = floor;
counter[index] += 1;
max = std::max(counter[index], max);
}
else
{
floor = std::max(max, floor);
}
}
for(std::vector<int>::iterator i = counter.begin();i != counter.end(); i++)
{
if(*i < floor)
*i = floor;
}
return counter;
}

Hera is my AC Java solution. The idea is the same as #Inwvr explained:
public int[] solution(int N, int[] A) {
int[] count = new int[N];
int max = 0;
int lastUpdate = 0;
for(int i = 0; i < A.length; i++){
if(A[i] <= N){
if(count[A[i]-1] < lastUpdate){
count[A[i]-1] = lastUpdate+1;
}
else{
count[A[i]-1]++;
}
max = Math.max(max, count[A[i]-1]);
}
else{
lastUpdate = max;
}
}
for(int i = 0; i < N; i++){
if(count[i] < lastUpdate)
count[i] = lastUpdate;
}
return count;
}

I just got 100 in PHP with some help from the above
function solution($N, $A) {
$B = array(0);
$max = 0;
foreach($A as $key => $a) {
$a -= 1;
if($a == $N) {
$max = max($B);
} else {
if(!isset($B[$a])) {
$B[$a] = 0;
}
if($B[$a] < $max) {
$B[$a] = $max + 1;
} else {
$B[$a] ++;
}
}
}
for($i=0; $i<$N; $i++) {
if(!isset($B[$i]) || $B[$i] < $max) {
$B[$i] = $max;
}
}
return $B;
}

This is another C++ solution to the problem.
The rationale is always the same.
Avoid to set to max counter all the counter upon instruction two, as this would bring the complexity to O(N*M).
Wait until we get another operation code on a single counter.
At this point the algorithm remembers whether it had met a max_counter and set the counter value consequently.
Here the code:
vector<int> MaxCounters(int N, vector<int> &A)
{
vector<int> n(N, 0);
int globalMax = 0;
int localMax = 0;
for( vector<int>::const_iterator it = A.begin(); it != A.end(); ++it)
{
if ( *it >= 1 && *it <= N)
{
// this is an increase op.
int value = *it - 1;
n[value] = std::max(n[value], localMax ) + 1;
globalMax = std::max(n[value], globalMax);
}
else
{
// set max counter op.
localMax = globalMax;
}
}
for( vector<int>::iterator it = n.begin(); it != n.end(); ++it)
*it = std::max( *it, localMax );
return n;
}

100%, O(m+n)
public int[] solution(int N, int[] A) {
int[] counters = new int[N];
int maxAIs = 0;
int minAShouldBe = 0;
for(int x : A) {
if(x >= 1 && x <= N) {
if(counters[x-1] < minAShouldBe) {
counters[x-1] = minAShouldBe;
}
counters[x-1]++;
if(counters[x-1] > maxAIs) {
maxAIs = counters[x-1];
}
} else if(x == N+1) {
minAShouldBe = maxAIs;
}
}
for(int i = 0; i < N; i++) {
if(counters[i] < minAShouldBe) {
counters[i] = minAShouldBe;
}
}
return counters;
}

here is my code, but its 88% cause it takes 3.80 sec for 10000 elements instead of 2.20
class Solution {
boolean maxCalled;
public int[] solution(int N, int[] A) {
int max =0;
int [] counters = new int [N];
int temp=0;
int currentVal = 0;
for(int i=0;i<A.length;i++){
currentVal = A[i];
if(currentVal <=N){
temp = increas(counters,currentVal);
if(temp > max){
max = temp;
}
}else{
if(!maxCalled)
maxCounter(counters,max);
}
}
return counters;
}
int increas (int [] A, int x){
maxCalled = false;
return ++A[x-1];
//return t;
}
void maxCounter (int [] A, int x){
maxCalled = true;
for (int i = 0; i < A.length; i++) {
A[i] = x;
}
}
}

Following my solution in JAVA (100/100).
public boolean isToSum(int value, int N) {
return value >= 1 && value <= N;
}
public int[] solution(int N, int[] A) {
int[] res = new int[N];
int max =0;
int minValue = 0;
for (int i=0; i < A.length; i++){
int value = A[i];
int pos = value -1;
if ( isToSum(value, N)) {
if( res[pos] < minValue) {
res[pos] = minValue;
}
res[pos] += 1;
if (max < res[pos]) {
max = res[pos];
}
} else {
minValue = max;
}
}
for (int i=0; i < res.length; i++){
if ( res[i] < minValue ){
res[i] = minValue;
}
}
return res;
}

my solution is :
public class Solution {
public int[] solution(int N, int[] A) {
int[] counters = new int[N];
int[] countersLastMaxIndexes = new int[N];
int maxValue = 0;
int fixedMaxValue = 0;
int maxIndex = 0;
for (int i = 0; i < A.length; i++) {
if (A[i] <= N) {
if (countersLastMaxIndexes[A[i] - 1] != maxIndex) {
counters[A[i] - 1] = fixedMaxValue;
countersLastMaxIndexes[A[i] - 1] = maxIndex;
}
counters[A[i] - 1]++;
if (counters[A[i] - 1] > maxValue) {
maxValue = counters[A[i] - 1];
}
} else {
maxIndex = i;
fixedMaxValue = maxValue;
}
}
for (int i = 0; i < countersLastMaxIndexes.length; i++) {
if (countersLastMaxIndexes[i] != maxIndex) {
counters[i] = fixedMaxValue;
countersLastMaxIndexes[i] = maxIndex;
}
}
return counters;
}
}

In my Java solution I updated values in solution[] only when needed. And finally updated solution[] with a right values.
public int[] solution(int N, int[] A) {
int[] solution = new int[N];
int maxCounter = 0;
int maxCountersSum = 0;
for(int a: A) {
if(a >= 1 && a <= N) {
if(solution[a - 1] < maxCountersSum)
solution[a - 1] = maxCountersSum;
solution[a - 1]++;
if(solution[a - 1] > maxCounter)
maxCounter = solution[a - 1];
}
if(a == N + 1) {
maxCountersSum = maxCounter;
}
}
for(int i = 0; i < N; i++) {
if(solution[i] < maxCountersSum)
solution[i] = maxCountersSum;
}
return solution;
}

Here's my python solution:
def solution(N, A):
# write your code in Python 3.6
RESP = [0] * N
MAX_OPERATION = N + 1
current_max = 0
current_min = 0
for operation in A:
if operation != MAX_OPERATION:
if RESP[operation-1] <= current_min:
RESP[operation-1] = current_min + 1
else:
RESP[operation-1] += 1
if RESP[operation-1] > current_max:
current_max = RESP[operation-1]
else:
if current_min == current_max:
current_min += 1
else:
current_min = current_max
for i, val in enumerate(RESP):
if val < current_min:
RESP[i] = current_min
return RESP

def sample_method(A,N=5):
initial_array = [0,0,0,0,0]
for i in A:
if(i>=1):
if(i<=N):
initial_array[i-1]+=1
else:
for a in range(len(initial_array)):
initial_array[a]+=1
print i
print initial_array

Here's my solution using python 3.6. The result is 100% correctness but 40% performance (most of them were because of timeout). Still cannot figure out how to optimize this code but hopefully someone can find it useful.
def solution(N, A):
count = [0]*(N+1)
for i in range(0,len(A)):
if A[i] >=1 and A[i] <= N:
count[A[i]] += 1
elif A[i] == (N+1):
count = [max(count)] * len(count)
count.pop(0)
return count

Typescript:
function counters(numCounters: number, operations: number[]) {
const counters = Array(numCounters)
let max = 0
let currentMin = 0
for (const operation of operations) {
if (operation === numCounters + 1) {
currentMin = max
} else {
if (!counters[operation - 1] || counters[operation - 1] < currentMin) {
counters[operation - 1] = currentMin
}
counters[operation - 1] = counters[operation - 1] + 1
if (counters[operation - 1] > max) {
max += 1
}
}
}
for (let i = 0; i < numCounters; i++) {
if (!counters[i] || counters[i] < currentMin) {
counters[i] = currentMin
}
}
return counters
}
console.log(solution=${counters(5, [3, 4, 4, 6, 1, 4, 4])})

100 points JavaScript solution, includes performance improvement to ignore repeated max_counter iterations:
function solution(N, A) {
let max = 0;
let counters = Array(N).fill(max);
let maxCounter = 0;
for (let op of A) {
if (op <= N && op >= 1) {
maxCounter = 0;
if (++counters[op - 1] > max) {
max = counters[op - 1];
}
} else if(op === N + 1 && maxCounter === 0) {
maxCounter = 1;
for (let i = 0; i < counters.length; i++) {
counters[i] = max;
}
}
}
return counters;
}

solution in JAVA (100/100)
class Solution {
public int[] solution(int N, int[] A) {
// write your code in Java SE 8
int[] result = new int[N];
int base = 0;
int max = 0;
int needToChange=A.length;;
for (int k = 0; k < A.length; k++) {
int X = A[k];
if (X >= 1 && X <= N) {
if (result[X - 1] < base) {
result[X - 1] = base;
}
result[X - 1]++;
if (max < result[X - 1]) {
max = result[X - 1];
}
}
if (X == N + 1) {
base = max;
needToChange= X-1;
}
}
for (int i = 0; i < needToChange; i++) {
if (result[i] < base) {
result[i] = base;
}
}
return result;
}
}

My Java solution. It gives 100% but is very long (in comparison). I have used HashMap for storing counters.
Detected time complexity: O(N + M)
import java.util.*;
class Solution {
final private Map<Integer, Integer> counters = new HashMap<>();
private int maxCounterValue = 0;
private int maxCounterValueRealized = 0;
public int[] solution(int N, int[] A) {
if (N < 1) return new int[0];
for (int a : A) {
if (a <= N) {
Integer current = counters.putIfAbsent(a, maxCounterValueRealized + 1);
if (current == null) {
updateMaxCounterValue(maxCounterValueRealized + 1);
} else {
++current;
counters.replace(a, current);
updateMaxCounterValue(current);
}
} else {
maxCounterValueRealized = maxCounterValue;
counters.clear();
}
}
return getCountersArray(N);
}
private void updateMaxCounterValue(int currentCounterValue) {
if (currentCounterValue > maxCounterValue)
maxCounterValue = currentCounterValue;
}
private int[] getCountersArray(int N) {
int[] countersArray = new int[N];
for (int j = 0; j < N; j++) {
Integer current = counters.get(j + 1);
if (current == null) {
countersArray[j] = maxCounterValueRealized;
} else {
countersArray[j] = current;
}
}
return countersArray;
}
}

Here is solution in python with 100 %
Codility Max counter 100%
def solution(N, A):
"""
Solution at 100% - https://app.codility.com/demo/results/trainingUQ95SB-4GA/
Idea is first take the counter array of given size N
take item from main A one by one + 1 and put in counter array , use item as index
keep track of last max operation
at the end replace counter items with max of local or counter item it self
:param N:
:param A:
:return:
"""
global_max = 0
local_max = 0
# counter array
counter = [0] * N
for i, item in enumerate(A):
# take item from original array one by one - 1 - minus due to using item as index
item_as_counter_index = item - 1
# print(item_as_counter_index)
# print(counter)
# print(local_max)
# current element less or equal value in array and greater than 1
# if A[K] = X, such that 1 ≤ X ≤ N, then operation K is increase(X),
if N >= item >= 1:
# max of local_max counter at item_as_counter_index
# increase counter array value and put in counter array
counter[item_as_counter_index] = max(local_max, counter[item_as_counter_index]) + 1
# track the status of global_max counter so far
# this is operation K
global_max = max(global_max, counter[item_as_counter_index])
# if A[K] = N + 1 then operation K is max counter.
elif item == N + 1:
# now operation k is as local max
# here we need to replace all items in array with this global max
# we can do using for loop for array length but that will cost bigo n2 complexity
# example - for i, item in A: counter[i] = global_max
local_max = global_max
# print("global_max each step")
# print(global_max)
# print("local max so far....")
# print(local_max)
# print("counter - ")
# print(counter)
# now counter array - replace all elements which are less than the local max found so far
# all counters are set to the maximum value of any counter
for i, item in enumerate(counter):
counter[i] = max(item, local_max)
return counter
result = solution(1, [3, 4, 4, 6, 1, 4, 4])
print("Sol " + str(result))

enter link description here
Got 100% result with O ( N + M )
class Solution {
public int[] solution(int N, int[] A) {
// write your code in Java SE 8
int max = 0;
int[] counter = new int[N];
int upgrade = 0;
for ( int i = 0; i < A.length; i++ )
{
if ( A[i] <= N )
{
if ( upgrade > 0 && upgrade > counter[A[i] - 1 ] )
{
counter[A[i] - 1] = upgrade;
}
counter[A[i] - 1 ]++;
if ( counter[A[i] - 1 ] > max )
{
max = counter[A[i] - 1 ];
}
}
else
{
upgrade = max;
}
}
for ( int i = 0; i < N; i++ )
{
if ( counter[i] < upgrade)
{
counter[i] = upgrade;
}
}
return counter;
}
}

Java 100%/100%, no imports
public int[] solution(int N, int[] A) {
int[] counters = new int[N];
int currentMax = 0;
int sumOfMaxCounters = 0;
boolean justDoneMaxCounter = false;
for (int i = 0; i < A.length ; i++) {
if (A[i] <= N) {
justDoneMaxCounter = false;
counters[A[i]-1]++;
currentMax = currentMax < counters[A[i]-1] ? counters[A[i]-1] : currentMax;
}else if (!justDoneMaxCounter){
sumOfMaxCounters += currentMax;
currentMax = 0;
counters = new int[N];
justDoneMaxCounter = true;
}
}
for (int j = 0; j < counters.length; j++) {
counters[j] = counters[j] + sumOfMaxCounters;
}
return counters;
}

python solution: 100% 100%
def solution(N, A):
c = [0] * N
max_element = 0
base = 0
for item in A:
if item >= 1 and N >= item:
c[item-1] = max(c[item-1], base) + 1
max_element = max(c[item - 1], max_element)
elif item == N + 1:
base = max_element
for i in range(N):
c[i] = max (c[i], base)
return c
pass

Using applyMax to record max operations
Time complexity:
O(N + M)
class Solution {
public int[] solution(int N, int[] A) {
// write your code in Java SE 8
int max = 0, applyMax = 0;;
int[] result = new int[N];
for (int i = 0; i < A.length; ++i) {
int a = A[i];
if (a == N + 1) {
applyMax = max;
}
if (1 <= a && a <= N) {
result[A[i] - 1] = Math.max(applyMax, result[A[i] - 1]);
max = Math.max(max, ++result[A[i] - 1]);
}
}
for (int i = 0; i < N; ++i) {
if (result[i] < applyMax) {
result[i] = applyMax;
}
}
return result;
}
}

Programming Test - Codility - Dominator [closed]

Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 4 years ago.
Improve this question
I just had a codility problem give me a hard time and I'm still trying to figure out how the space and time complexity constraints could have been met.
The problem is as follows:
A dominant member in the array is one that occupies over half the positions in the array, for example:
{3, 67, 23, 67, 67}
67 is a dominant member because it appears in the array in 3/5 (>50%) positions.
Now, you are expected to provide a method that takes in an array and returns an index of the dominant member if one exists and -1 if there is none.
Easy, right? Well, I could have solved the problem handily if it were not for the following constraints:
Expected time complexity is O(n)
Expected space complexity is O(1)
I can see how you could solve this for O(n) time with O(n) space complexities as well as O(n^2) time with O(1) space complexities, but not one that meets both O(n) time and O(1) space.
I would really appreciate seeing a solution to this problem. Don't worry, the deadline has passed a few hours ago (I only had 30 minutes), so I'm not trying to cheat. Thanks.

Googled "computing dominant member of array", it was the first result. See the algorithm described on page 3.
element x;
int count ← 0;
For(i = 0 to n − 1) {
if(count == 0) { x ← A[i]; count++; }
else if (A[i] == x) count++;
else count−−;
}
Check if x is dominant element by scanning array A
Basically observe that if you find two different elements in the array, you can remove them both without changing the dominant element on the remainder. This code just keeps tossing out pairs of different elements, keeping track of the number of times it has seen the single remaining unpaired element.

Find the median with BFPRT, aka median of medians (O(N) time, O(1) space). Then scan through the array -- if one number dominates, the median will be equal to that number. Walk through the array and count the number of instances of that number. If it's over half the array, it's the dominator. Otherwise, there is no dominator.

Adding a Java 100/100 O(N) time with O(1) space:
https://codility.com/demo/results/demoPNG8BT-KEH/
class Solution {
public int solution(int[] A) {
int indexOfCandidate = -1;
int stackCounter = 0, candidate=-1, value=-1, i =0;
for(int element: A ) {
if (stackCounter == 0) {
value = element;
++stackCounter;
indexOfCandidate = i;
} else {
if (value == element) {
++stackCounter;
} else {
--stackCounter;
}
}
++i;
}
if (stackCounter > 0 ) {
candidate = value;
} else {
return -1;
}
int countRepetitions = 0;
for (int element: A) {
if( element == candidate) {
++countRepetitions;
}
if(countRepetitions > (A.length / 2)) {
return indexOfCandidate;
}
}
return -1;
}
}
If you want to see the Java source code it's here, I added some test cases as comments as the beginning of the file.

Java solution with score 100%
public int solution(int[] array) {
int candidate=0;
int counter = 0;
// Find candidate for leader
for(int i=0; i<array.length; i++){
if(counter == 0) candidate = i;
if(array[i] == array[candidate]){
counter++;
}else {
counter--;
}
}
// Count candidate occurrences in array
counter = 0;
for(int i=0; i<array.length; i++){
if(array[i] == array[candidate]) counter++;
}
// Check that candidate occurs more than array.lenght/2
return counter>array.length/2 ? candidate : -1;
}

In python, we are lucky some smart people have bothered to implement efficient helpers using C and shipped it in the standard library. The collections.Counter is useful here.
>>> data = [3, 67, 23, 67, 67]
>>> from collections import Counter
>>> counter = Counter(data) # counter accepts any sequence/iterable
>>> counter # dict like object, where values are the occurrence
Counter({67: 3, 3: 1, 23: 1})
>>> common = counter.most_common()[0]
>>> common
(67, 3)
>>> common[0] if common[1] > len(data) / 2.0 + 1 else -1
67
>>>
If you prefer a function here is one ...
>>> def dominator(seq):
counter = Counter(seq)
common = counter.most_common()[0]
return common[0] if common[1] > len(seq) / 2.0 + 1 else -1
...
>>> dominator([1, 3, 6, 7, 6, 8, 6])
-1
>>> dominator([1, 3, 6, 7, 6, 8, 6, 6])
6

This question looks hard if a small trick does not come to the mind :). I found this trick in this document of codility : https://codility.com/media/train/6-Leader.pdf.
The linear solution is explained at the bottom of this document.
I implemented the following java program which gave me a score of 100 on the same lines.
public int solution(int[] A) {
Stack<Integer> stack = new Stack<Integer>();
for (int i =0; i < A.length; i++)
{
if (stack.empty())
stack.push(new Integer(A[i]));
else
{
int topElem = stack.peek().intValue();
if (topElem == A[i])
{
stack.push(new Integer(A[i]));
}
else
{
stack.pop();
}
}
}
if (stack.empty())
return -1;
int elem = stack.peek().intValue();
int count = 0;
int index = 0;
for (int i = 0; i < A.length; i++)
{
if (elem == A[i])
{
count++;
index = i;
}
}
if (count > ((double)A.length/2.0))
return index;
else
return -1;
}

Here's my C solution which scores 100%
int solution(int A[], int N) {
int candidate;
int count = 0;
int i;
// 1. Find most likely candidate for the leader
for(i = 0; i < N; i++){
// change candidate when count reaches 0
if(count == 0) candidate = i;
// count occurrences of candidate
if(A[i] == A[candidate]) count++;
else count--;
}
// 2. Verify that candidate occurs more than N/2 times
count = 0;
for(i = 0; i < N; i++) if(A[i] == A[candidate]) count++;
if (count <= N/2) return -1;
return candidate; // return index of leader
}

100%
import java.util.HashMap;
import java.util.Map;
class Solution {
public static int solution(int[] A) {
final int N = A.length;
Map<Integer, Integer> mapOfOccur = new HashMap((N/2)+1);
for(int i=0; i<N; i++){
Integer count = mapOfOccur.get(A[i]);
if(count == null){
count = 1;
mapOfOccur.put(A[i],count);
}else{
mapOfOccur.replace(A[i], count, ++count);
}
if(count > N/2)
return i;
}
return -1;
}
}

Does it have to be a particularly good algorithm? ;-)
static int dominant(final int... set) {
final int[] freqs = new int[Integer.MAX_VALUE];
for (int n : set) {
++freqs[n];
}
int dom_freq = Integer.MIN_VALUE;
int dom_idx = -1;
int dom_n = -1;
for (int i = set.length - 1; i >= 0; --i) {
final int n = set[i];
if (dom_n != n) {
final int freq = freqs[n];
if (freq > dom_freq) {
dom_freq = freq;
dom_n = n;
dom_idx = i;
} else if (freq == dom_freq) {
dom_idx = -1;
}
}
}
return dom_idx;
}
(this was primarily meant to poke fun at the requirements)

Consider this 100/100 solution in Ruby:
# Algorithm, as described in https://codility.com/media/train/6-Leader.pdf:
#
# * Iterate once to find a candidate for dominator.
# * Count number of candidate occurences for the final conclusion.
def solution(ar)
n_occu = 0
candidate = index = nil
ar.each_with_index do |elem, i|
if n_occu < 1
# Here comes a new dominator candidate.
candidate = elem
index = i
n_occu += 1
else
if candidate == elem
n_occu += 1
else
n_occu -= 1
end
end # if n_occu < 1
end
# Method result. -1 if no dominator.
# Count number of occurences to check if candidate is really a dominator.
if n_occu > 0 and ar.count {|_| _ == candidate} > ar.size/2
index
else
-1
end
end
#--------------------------------------- Tests
def test
sets = []
sets << ["4666688", [1, 2, 3, 4], [4, 6, 6, 6, 6, 8, 8]]
sets << ["333311", [0, 1, 2, 3], [3, 3, 3, 3, 1, 1]]
sets << ["313131", [-1], [3, 1, 3, 1, 3, 1]]
sets << ["113333", [2, 3, 4, 5], [1, 1, 3, 3, 3, 3]]
sets.each do |name, one_of_expected, ar|
out = solution(ar)
raise "FAILURE at test #{name.inspect}: #{out.inspect} not in #{expected.inspect}" if not one_of_expected.include? out
end
puts "SUCCESS: All tests passed"
end

Here is an easy to read, 100% score version in Objective-c
if (A.count > 100000)
return -1;
NSInteger occur = 0;
NSNumber *candidate = nil;
for (NSNumber *element in A){
if (!candidate){
candidate = element;
occur = 1;
continue;
}
if ([candidate isEqualToNumber:element]){
occur++;
}else{
if (occur == 1){
candidate = element;
continue;
}else{
occur--;
}
}
}
if (candidate){
occur = 0;
for (NSNumber *element in A){
if ([candidate isEqualToNumber:element])
occur++;
}
if (occur > A.count / 2)
return [A indexOfObject:candidate];
}
return -1;

100% score JavaScript solution. Technically it's O(nlogn) but still passed.
function solution(A) {
if (A.length == 0)
return -1;
var S = A.slice(0).sort(function(a, b) {
return a - b;
});
var domThresh = A.length/2;
var c = S[Math.floor(domThresh)];
var domCount = 0;
for (var i = 0; i < A.length; i++) {
if (A[i] == c)
domCount++;
if (domCount > domThresh)
return i;
}
return -1;
}

This is the solution in VB.NET with 100% performance.
Dim result As Integer = 0
Dim i, ladderVal, LadderCount, size, valCount As Integer
ladderVal = 0
LadderCount = 0
size = A.Length
If size > 0 Then
For i = 1 To size - 1
If LadderCount = 0 Then
LadderCount += 1
ladderVal = A(i)
Else
If A(i) = ladderVal Then
LadderCount += 1
Else
LadderCount -= 1
End If
End If
Next
valCount = 0
For i = 0 To size - 1
If A(i) = ladderVal Then
valCount += 1
End If
Next
If valCount <= size / 2 Then
result = 0
Else
LadderCount = 0
For i = 0 To size - 1
If A(i) = ladderVal Then
valCount -= 1
LadderCount += 1
End If
If LadderCount > (LadderCount + 1) / 2 And (valCount > (size - (i + 1)) / 2) Then
result += 1
End If
Next
End If
End If
Return result
See the correctness and performance of the code

Below solution resolves in complexity O(N).
public int solution(int A[]){
int dominatorValue=-1;
if(A != null && A.length>0){
Hashtable<Integer, Integer> count=new Hashtable<>();
dominatorValue=A[0];
int big=0;
for (int i = 0; i < A.length; i++) {
int value=0;
try{
value=count.get(A[i]);
value++;
}catch(Exception e){
}
count.put(A[i], value);
if(value>big){
big=value;
dominatorValue=A[i];
}
}
}
return dominatorValue;
}

100% in PHP https://codility.com/demo/results/trainingVRQGQ9-NJP/
function solution($A){
if (empty($A)) return -1;
$copy = array_count_values($A); // 3 => 7, value => number of repetition
$max_repetition = max($copy); // at least 1 because the array is not empty
$dominator = array_search($max_repetition, $copy);
if ($max_repetition > count($A) / 2) return array_search($dominator, $A); else return -1;
}

i test my code its work fine in arrays lengths between 2 to 9
public static int sol (int []a)
{
int count = 0 ;
int candidateIndex = -1;
for (int i = 0; i <a.length ; i++)
{
int nextIndex = 0;
int nextOfNextIndex = 0;
if(i<a.length-2)
{
nextIndex = i+1;
nextOfNextIndex = i+2;
}
if(count==0)
{
candidateIndex = i;
}
if(a[candidateIndex]== a[nextIndex])
{
count++;
}
if (a[candidateIndex]==a[nextOfNextIndex])
{
count++;
}
}
count -- ;
return count>a.length/2?candidateIndex:-1;
}

Adding a Java 100/100 O(N) time with O(1) space:
// you can also use imports, for example:
import java.util.Stack;
// you can write to stdout for debugging purposes, e.g.
// System.out.println("this is a debug message");
class Solution {
public int solution(int[] A) {
// write your code in Java SE 8
int count = 0;
Stack<Integer> integerStack = new Stack<Integer>();
for (int i = 0; i < A.length; i++) {
if (integerStack.isEmpty()) {
integerStack.push(A[i]);
} else if (integerStack.size() > 0) {
if (integerStack.peek() == A[i])
integerStack.push(A[i]);
else
integerStack.pop();
}
}
if (!integerStack.isEmpty()) {
for (int i = 0; i < integerStack.size(); i++) {
for (int j = 0; j < A.length; j++) {
if (integerStack.get(i) == A[j])
count++;
if (count > A.length / 2)
return j;
}
count = 0;
}
}
return -1;
}
}
Here is test result from codility.

I think this question has already been resolved somewhere. The "official" solution should be :
public int dominator(int[] A) {
int N = A.length;
for(int i = 0; i< N/2+1; i++)
{
int count=1;
for(int j = i+1; j < N; j++)
{
if (A[i]==A[j]) {count++; if (count > (N/2)) return i;}
}
}
return -1;
}

How about sorting the array first? You then compare middle and first and last elements of the sorted array to find the dominant element.
public Integer findDominator(int[] arr) {
int[] arrCopy = arr.clone();
Arrays.sort(arrCopy);
int length = arrCopy.length;
int middleIndx = (length - 1) /2;
int middleIdxRight;
int middleIdxLeft = middleIndx;
if (length % 2 == 0) {
middleIdxRight = middleIndx+1;
} else {
middleIdxRight = middleIndx;
}
if (arrCopy[0] == arrCopy[middleIdxRight]) {
return arrCopy[0];
}
if (arrCopy[middleIdxLeft] == arrCopy[length -1]) {
return arrCopy[middleIdxLeft];
}
return null;
}

C#
int dominant = 0;
int repeat = 0;
int? repeatedNr = null;
int maxLenght = A.Length;
int halfLenght = A.Length / 2;
int[] repeations = new int[A.Length];
for (int i = 0; i < A.Length; i++)
{
repeatedNr = A[i];
for (int j = 0; j < A.Length; j++)
{
if (repeatedNr == A[j])
{
repeations[i]++;
}
}
}
repeatedNr = null;
for (int i = 0; i < repeations.Length; i++)
{
if (repeations[i] > repeat)
{
repeat = repeations[i];
repeatedNr = A[i];
}
}
if (repeat > halfLenght)
dominant = int.Parse(repeatedNr.ToString());

class Program
{
static void Main(string[] args)
{
int []A= new int[] {3,6,2,6};
int[] B = new int[A.Length];
Program obj = new Program();
obj.ABC(A,B);
}
public int ABC(int []A, int []B)
{
int i,j;
int n= A.Length;
for (j=0; j<n ;j++)
{
int count = 1;
for (i = 0; i < n; i++)
{
if ((A[j]== A[i] && i!=j))
{
count++;
}
}
int finalCount = count;
B[j] = finalCount;// to store the no of times a number is repeated
}
// int finalCount = count / 2;
int finalCount1 = B.Max();// see which number occurred max times
if (finalCount1 > (n / 2))
{ Console.WriteLine(finalCount1); Console.ReadLine(); }
else
{ Console.WriteLine("no number found"); Console.ReadLine(); }
return -1;
}
}

In Ruby you can do something like
def dominant(a)
hash = {}
0.upto(a.length) do |index|
element = a[index]
hash[element] = (hash[element] ? hash[element] + 1 : 1)
end
res = hash.find{|k,v| v > a.length / 2}.first rescue nil
res ||= -1
return res
end

#Keith Randall solution is not working for {1,1,2,2,3,2,2}
his solution was:
element x;
int count ← 0;
For(i = 0 to n − 1) {
if(count == 0) { x ← A[i]; count++; }
else if (A[i] == x) count++;
else count−−;
}
Check if x is dominant element by scanning array A
I converted it into java as below:
int x = 0;
int count = 0;
for(int i = 0; i < (arr.length - 1); i++) {
if(count == 0) {
x = arr[i];
count++;
}
else if (arr[i] == x)
count++;
else count--;
}
return x;
Out put : 3
Expected: 2

This is my answer in Java: I store a count in seperate array which counts duplicates of each of the entries of the input array and then keeps a pointer to the array position that has the most duplicates. This is the dominator.
private static void dom(int[] a) {
int position = 0;
int max = 0;
int score = 0;
int counter = 0;
int[]result = new int[a.length];
for(int i = 0; i < a.length; i++){
score = 0;
for(int c = 0; c < a.length;c++){
if(a[i] == a[c] && c != i ){
score = score + 1;
result[i] = score;
if(result[i] > position){
position = i;
}
}
}
}
//This is just to facilitate the print function and MAX = the number of times that dominator number was found in the list.
for(int x = 0 ; x < result.length-1; x++){
if(result[x] > max){
max = result[x] + 1;
}
}
System.out.println(" The following number is the dominator " + a[position] + " it appears a total of " + max);
}