We Keep Coding

Rearrange elements in an integer array in a fashion where-you first replace and then move the replaced integer to first part of the array

I came across a Hackerearth coding problem where you have to perform the following tasks over an integer array-
Search for a particular number in the array and replace it's occurrences with 1
Move all the 1s to the first part of the array, maintaining the original order of the array
For example- if we have an integer array {22,1,34,22,16,22,35,1}, here we search for the number "22" (let us assume it is present in the array), replace it with 1 and move all those 1s (including the 1s already present) to the first part of the array and the resultant array should look like {1,1,1,1,1,1,34,16,35} -maintaining the original order of the array, preferably in Java.
I actually have coded a solution and it works fine but is not optimal, can anyone help me find an optimal solution (w.r.t. time-space complexity)?
Below is my solution-
public static void main(String[] args) {
int[] n = rearr(new int[] {22,1,34,22,16,22,1,34,1}, 22);
for(int i=0; i<n.length; i++) {
System.out.print(n[i]+" ");
}
}
static int[] rearr(int[] a, int x) {
int[] temp = new int[a.length];
int j=0, c=0, k=0;
//search and replace
for(int i=0; i<a.length; i++) {
if(a[i] == x) {
a[i] = 1;
}
}
//shift all 1s to first part of array or shift all non-1s to last part of the array
for(int i=0; i<a.length; i++) {
if(a[i] != 1) {
temp[j] = a[i];
j++;
}
if(a[i] == 1) {
c++;
}
}
j=0;
for(int i=0; i<a.length && c>0; i++, c--) {
a[i] = 1;
j++;
}
for(int i=j ;i<a.length; i++) {
a[i] = temp[k];
k++;
}
return a;
}

This can be done in linear time and space complexity, by returning a completely new list instead of modifying the original list.
static int[] rearr(int[] a, int x) {
// allocate the array we'll return
int[] b = new int[a.length];
int fillvalue = 1;
// iterate backwards through the list, and transplant every value OTHER than
// (x or 1) to the last open index in b, which we track with b_idx
int b_idx = b.length - 1;
for (int i = a.length - 1; i >= 0; i--) {
if (a[i] != x && a[i] != fillvalue)) {
b[b_idx] = a[i];
b_idx--;
}
}
// once we've gone through and done that, fill what remains of b with ones
// which are either original or are replacements, we don't care
for (int i = b_idx; i >= 0; i--) {
b[i] = fillvalue;
}
return b;
}
This is linear space complexity because it requires additional space equal to the size of the given list. It's linear time complexity because, in the worst case, it iterates over the size of the list exactly twice.
As a bonus, if we decide we want to leave the original 1s where they were, we can do that without any trouble at all, by simply modifying the if condition. Same if we decide we want to change the fill value to something else.
Doing this with constant space complexity would require O(n^2) list complexity, as it would require swapping elements in a to their proper positions. The easiest way to do that would probably be to do replacements on a first run through the list, and then do something like bubblesort to move all the 1s to the front.

This can be done in a single iteration through the array. We can use 2 pointer approach here where we will use on pointer to iterate through the array and other one to point to the index of 1 in the array.
The code is below:
public static void main(String[] args) {
// input array
int[] arr = { 22, 1, 34, 22, 16, 22, 35, 1, 20, 33, 136 };
// element to be replaced
int x = 22;
int j = -1;
for (int i = arr.length - 1; i >= 0; i--) {
if (arr[i] == 1 || arr[i] == x) {
if (j == -1) {
j = i;
}
// incase arr[i]==x
arr[i] = 1;
} else {
if (j != -1) {
arr[j] = arr[i];
arr[i] = 1;
j--;
}
}
}
for (int i = 0; i < arr.length; i++) {
System.out.println(arr[i]);
}
}
Here we initialise j=-1 since we consider there are no 1's present in the array.
Then we start iterating the array from the end towards the starting of the array as we have to push all the 1's to the starting of the array. Now when we reach to 1 or x (particular number in your case), we check if this is first occurrence of the x or 1, if yes then we initialise the j with this index and change arr[i] = 1 because this could be equal to x then we need to make it 1. If the arr[i] is not 1 or x it means its a number which we need to push at back of the array. We check if we have position of 1 or j=-1. If j=-1 it means this number is already pushed back at end of array else we swap the number at i and j, and decrement j by 1.
At the end of the array we will have the array sorted in a fashion which is required.
Time Complexity: Since we are only iterating the array one, hence the time complexity is O(n).
Space Complexity: Since there are no extra space being used or constant space being used hence the space complexity is O(1)

Trying to create a array with the intersection of two arrays but fails at creating array with the proper structure

So, I am trying to create 2 randomly generated arrays,(a, and b, each with 10 unique whole numbers from 0 to 20), and then creating 2 arrays with the info of the last two. One containing the numbers that appear in both a and b, and another with the numbers that are unique to a and to b. The arrays must be listed in a "a -> [1, 2, 3,...]" format. At the moment I only know how to generate the 2 arrays, and am currently at the Intersection part. The problem is, that I can create a array with the correct list of numbers, but it will have the same length of the other two, and the spaces where it shouldn't have anything, it will be filled with 0s when its supposed to create a smaller array with only the right numbers.
package tps.tp1.pack2Arrays;
public class P02ArraysExtractUniqsAndReps {
public static void main(String[] args) {
int nbr = 10;
int min = 0;
int max = 20;
generateArray(nbr, min, max);
System.out.println();
}
public static int[] generateArray(int nbr, int min, int max) {
int[] a = new int[nbr];
int[] b = new int[nbr];
int[] s = new int[nbr];
s[0] = 0;
for (int i = 0; i < a.length; i++) {
a[i] = (int) (Math.random() * (max - min));
b[i] = (int) (Math.random() * (max - min));
for (int j = 0; j < i; j++) {
if (a[i] == a[j]) {
i--;
}
if (b[i] == b[j]) {
i--;
}
}
}
System.out.println("a - > " + Arrays.toString(a));
System.out.println("b - > " + Arrays.toString(b));
for (int k = 0; k < a.length; k++) {
for (int l = 0; l < b.length; l++) {
if (a[k] == b[l]) {
s[l] = b[l];
}else {
}
}
}
System.out.println("(a ∪ (b/(a ∩ b)) - > " + Arrays.toString(s));
return null;
}
public static boolean hasValue(int[] array, int value) {
for (int i = 0; i < array.length; i++) {
if (array[i] == value) {
return true;
}
}
return false;
}
}
Is there any way to create the array without the incorrect 0s? (I say incorrect because it is possible to have 0 in both a and b).
Any help/clarification is appreciated.

First, allocate an array large enough to hold the intersection. It needs to be no bigger that the smaller of the source arrays.
When you add a value to the intersection array, always add it starting at the beginning of the array. Use a counter to update the next position. This also allows the value 0 to be a valid value.
Then when finished. use Array.copyOf() to copy only the first part of the array to itself, thus removing the empty (unfilled 0 value) spaces. This works as follow assuming count is the index you have been using to add to the array: Assume count = 3
int[] inter = {1,2,3,0,0,0,0};
inter = Arrays.copyOf(inter, count);
System.out.println(Arrays.toString(inter);
prints
[1,2,3]
Here is an approach using a List
int[] b = {4,3,1,2,5,0,2};
int [] a = {3,5,2,3,7,8,2,0,9,10};
Add one of the arrays to the list.
List<Integer> list = new ArrayList<>();
for(int i : a) {
list.add(i);
}
Allocate the intersection array with count used as the next location. It doesn't matter which array's length you use.
int count = 0;
int [] intersection = new int[a.length];
Now simply iterate thru the other array.
if the list contains the value, add it to the intersection array.
then remove it from the list and increment count. NOTE - The removed value must be converted to an Integer object, otherwise, if a simple int value, it would be interpreted as an index and the value at that index would be removed and not the actual value itself (or an Exception might be thrown).
once finished the intersection array will have the values and probably unseen zeroes at the end.
for(int i = 0; i < b.length; i++) {
int val = b[i];
if (list.contains(val)) {
intersection[count++] = val;
list.remove(Integer.valueOf(val));
}
}
To shorten the array, use the copy method mentioned above.
intersection = Arrays.copyOf(intersection, count);
System.out.println(Arrays.toString(intersection));
prints
[3, 2, 5, 0, 2]
Note that it does not matter which array is which. If you reverse the arrays for a and b above, the same intersection will result, albeit in a different order.

The first thing I notice is that you are declaring your intersection array at the top of the method.
int[] s = new int[nbr];
You are declaring the same amount of space for the array regardless of the amount you actually use.
Method Arrays.toString(int []) will print any uninitialized slots in the array as "0"
There are several different approaches you can take here:
You can delay initializing the array until you have determined the size of the set you are dealing with.
You can transfer your content into another well sized array after figuring out your result set.
You could forego using Array.toString, and build the string up yourself.

CodeFights - issues with time limit when writing FirstDuplicate method

I'm trying to solve the problem below from CodeFights. I left my answer in Java after the question. The code works for all the problems, except the last one. Time limit exception is reported. What could I do to make it run below 3000ms (CodeFights requirement)?
Note: Write a solution with O(n) time complexity and O(1) additional space complexity, since this is what you would be asked to do during a real interview.
Given an array a that contains only numbers in the range from 1 to a.length, find the first duplicate number for which the second occurrence has the minimal index. In other words, if there are more than 1 duplicated numbers, return the number for which the second occurrence has a smaller index than the second occurrence of the other number does. If there are no such elements, return -1.
Example
For a = [2, 3, 3, 1, 5, 2], the output should be
firstDuplicate(a) = 3.
There are 2 duplicates: numbers 2 and 3. The second occurrence of 3 has a smaller index than than second occurrence of 2 does, so the answer is 3.
For a = [2, 4, 3, 5, 1], the output should be
firstDuplicate(a) = -1.
Input/Output
[time limit] 3000ms (java)
[input] array.integer a
Guaranteed constraints:
1 ≤ a.length ≤ 105,
1 ≤ a[i] ≤ a.length.
[output] integer
The element in a that occurs in the array more than once and has the minimal index for its second occurrence. If there are no such elements, return -1.
int storedLeastValue = -1;
int indexDistances = Integer.MAX_VALUE;
int indexPosition = Integer.MAX_VALUE;
for (int i = 0; i < a.length; i++)
{
int tempValue = a[i];
for (int j = i+1; j < a.length; j++) {
if(tempValue == a[j])
{
if(Math.abs(i-j) < indexDistances &&
j < indexPosition)
{
storedLeastValue = tempValue;
indexDistances = Math.abs(i-j);
indexPosition = j;
break;
}
}
}
}
return storedLeastValue;

Your solution has two nested for loops which implies O(n^2) while the question explicitly asks for O(n). Since you also have a space restriction you can't use an additional Set (which can provide a simple solution as well).
This question is good for people that have strong algorithms/graph theory background. The solution is sophisticated and includes finding an entry point for a cycle in a directed graph. If you're not familiar with these terms I'd recommend that you'll leave it and move to other questions.

Check this one, it's also O(n) , but without additional array.
int firstDuplicate(int[] a) {
if (a.length <= 1) return -1;
for (int i = 0; i < a.length; i++) {
int pos = Math.abs(a[i]) - 1;
if (a[pos] < 0) return pos + 1;
else a[pos] = -a[pos];
}
return -1;
}

The accepted answer does not work with the task.
It would work if the input array would indeed contain no bigger value than its length.
But it does, eg.: [5,5].
So, we have to define which number is the biggest.
int firstDuplicate(int[] a) {
int size = 0;
for(int i = 0; i < a.length; i++) {
if(a[i] > size) {
size = a[i];
}
}
int[] t = new int[size+1];
for(int i = 0; i < a.length; i++) {
if(t[a[i]] == 0) {
t[a[i]]++;
} else {
return a[i];
}
}
return -1;
}

What about this:
public static void main(String args[]) {
int [] a = new int[] {2, 3, 3, 1, 5, 2};
// Each element of cntarray will hold the number of occurrences of each potential number in the input (cntarray[n] = occurrences of n)
// Default initialization to zero's
int [] cntarray = new int[a.length + 1]; // need +1 in order to prevent index out of bounds errors, cntarray[0] is just an empty element
int min = -1;
for (int i=0;i < a.length ;i++) {
if (cntarray[a[i]] == 0) {
cntarray[a[i]]++;
} else {
min = a[i];
// no need to go further
break;
}
}
System.out.println(min);
}

You can store array values in hashSet. Check if value is already present in hashSet if not present then add it in hashSet else that will be your answer. Below is code which passes all test cases:-
int firstDuplicate(int[] a) {
HashSet<Integer> hashSet = new HashSet<>();
for(int i=0; i<a.length;i++){
if (! hashSet.contains(a[i])) {
hashSet.add(a[i]);
} else {
return a[i];
}
}
return -1;
}

My simple solution with a HashMap
int solution(int[] a) {
HashMap<Integer, Integer> countMap = new HashMap<Integer, Integer>();
int min = -1;
for (int i=0; i < a.length; i++) {
if (!(countMap.containsKey(a[i]))) {
countMap.put(a[i],1);
}
else {
return a[i];
}
}
return min;
}

Solution is very simple:
Create a hashset
keep iterating over the array
if element is already not in the set, add it.
else element will be in the set, then it mean this is minimal index of first/second the duplicate
int solution(int[] a) {
HashSet<Integer> set = new HashSet<>();
for(int i=0; i<a.length; i++){
if(set.contains(a[i])){
// as soon as minimal index duplicate found where first one was already in the set, return it
return a[i];
}
set.add(a[i]);
}
return -1;
}

A good answer for this exercise can be found here - https://forum.thecoders.org/t/an-interesting-coding-problem-in-codefights/163 - Everything is done in-place, and it has O(1) solution.

Count The Amount Of Data In An Array Including SOME Null

I'm coding in java and I need to create a function that returns the number of data objects that are currently in an ArrayList. At the moment I have this:
int count = 0;
for (int i = 0; i < data.length; i++)
{
if (data[i] != null)
{
count ++;
}
}
return count;
But the problem is that an array list that includes null data is acceptable, and I have to count their null data towards this counter. How do I include the null data that's in the middle of this array, and not the null data that's not supposed to be counted for?
For example, I have some tester code that adds (8),null,null,(23),(25) to the array, and this function should return 5 when the initial array size is 10.

I'm going to assume you're using a regular array (your question is somewhat ambiguous about this). Traverse through the array backwards until you find a non-null element:
public static int count(Object[] a) {
int i = a.length - 1;
for (; i >= 0 ; i--)
if (a[i] != null)
break;
return i + 1;
}
You could also have
public static <T> int count(T[] a) {
int i = a.length - 1;
for (; i >= 0 ; i--)
if (a[i] != null)
break;
return i + 1;
}
Let's test it out, using an example analogous to the one you provided:
Object[] a = new Object[10];
a[0] = new Object();
a[3] = new Object();
a[4] = new Object();
System.out.println(count(a));
Output:
5

You will need two separate counters. The first one will count normally. The second one starts counting when you find null data. Then when you find a non-null data, just add the second counter to the first one and continue counting with the first counter until you find a null again.

int count = 0;
for (int i = data.length - 1; i >= 0; i--)
if (data[i] != null || count > 0)
count += 1;
return count;
At least that's how I understood your requirements - count nulls, except for trailing nulls.
But maybe that's not actually what you meant?
Edit
Unless you're actually using ArrayList (as Jon was asking), where .size() is different from capacity and will count all added elements (including nulls). You can't actually even get the capacity from an ArrayList.

How to compare two arrays of integers order-insensitively

I want Java code that can compare in this way (for example):
<1 2 3 4> = <3 1 2 4>
<1 2 3 4> != <3 4 1 1>
I can't use hashmap table or anything; just pure code without library.
I know there are two ways.
sort them and compare the array index by index
use two for loops and compare the outer index with the inner index. I have been trying with this but still not working:
for(int i = 0; i < n; i++)
{
for(int j = 0; j < n; j++)
{
if(a[i] != a[j] && j == n)
return false;
}
}
return true;
anything wrong with the code ? thanks

Sort & compare. You can't get the complexity better than that, and any "improvement" that you do on the speed comes at the risk that your code will be wrong.
[edit] Actually.... if you know your numbers are relatively small (e.g. say: the arrays only contain numbers between 0 and 1000), that there's an alternative in O(n). Something like this (sorry if the syntax is wrong, I didn't use java lately):
int count[1001]; // already intialized to 0
for(int i=0;i<n;i++){ count[a[i]]++; count[b[i]]--;}
bool arrays_identical = true;
for(int i=0;i<=1000 && arrays_identical; i++)
arrays_identical &= count[i]==0;
Disclaimer: this code doesn't contain "sanity checks" (i.e. the arrays are really of length "n", the numbers are in the prescribed interval) - it's only to illustrate the principle.

Arrays.sort(a);
Arrays.sort(b);
return Arrays.equals(a, b);

Since this is homework, I'm going to guess that a simple quadratic solution (which your original attempt is), is fine. In that case, you can do something like this:
int count(int[] nums, int x) {
int count = 0;
for (int num : nums) {
if (num == x) count++;
}
return count;
}
boolean equals(int[] arr1, int[] arr2) {
if (arr1.length != arr2.length) return false;
for (int x : arr1) {
if (count(arr1, x) != count(arr2, x)) return false;
}
return true;
}
This sacrifices speed for clarity: it's obviously correct!!
Tips
Use for-each whenever applicable; it always read to better readability
Helper functions improve both readability and reusability!

Before you actually start a more computationally complex solution, you can run a fast O(n) test to see if the arrays are the same. There are times when this this will result in a false positive, and you can use more computationally intensive means to investigate further. If it returns false, you can be assured that the arrays are different.
The basic approach is to do a cummulative XOR between the ith elements of the two arrays. If the resulting XOR is non-zero then you know that the two arrays are different. If the XOR is zero, they might be the same, but more work is needed to confirm.
int c_xor = 0; // XOR accumulator
for (int i = 0; i < n; ++i)
c_xor ^= arr1[i] ^ arr2[i];
// They're different if c_xor != 0; the might be the same if c_xor == 0
return c_xor == 0;

I think it would be sensible to check that the lengths of both arrays are equal before doing any iterating. It's a cheap way to opt out of hard work early, and it fixes the following type of failure for the second approach you mention:
{1, 2, 3} seen as equal to {4, 3, 2, 1}

This is a O(n^2) problem. No other solution is faster than comparing the two arrays directly.
The Virgil's solution is cool, except for the fact that it is not really O(n) performance. It is really O(n+1000) performance. The array comparison second time to set the boolean variable is costly and backfires in small arrays.
The solution you wrote is the best except for a bugs.
Here is the error corrected version.
boolean[] matchedPositions = new boolean[n];
for(int k = 0; k < n; k++
{
matchedPositions[k] = false;
}
for(int i = 0; i < n; i++)
{
for(int j = 0; j < n; j++)
{
if(matchedPositions[j] == false && a[i] == a[j])
{
matchedPositions[j] = true;
break;
}
if(j == n - 1)
{
return false;
}
}
}
return true;
In this case if in case the integer matches then the inner loop will break and set or flag the matched position. This is required for arrays with duplicate entries, this will avoid two entries in left array matching with one entry in the right array.
If in case the match did not happen, which is identified by j == n - 1, you will return false.
Since we are expecting a default value of false in boolean its better to flag initialize it.
In reality, this solution has O(n log n + n) performance penalty. Sort and compare has a performance penalty of O(n^2 + n) too. Sort has O(n^2) performance and one loop for checking. But sorting changes the contents of the array and this does not.

if you take one array as 1,2,3,4,3,1,2,4 then the solution is in this way.
2n = total number of integers : 8
//program
int i, j, n = 4;
for(i = 0; i < n; i++)
for(j = n; j < 2n; j++)
{
if( a[i] != a[j])
{
j++;
}
else
{
i++; exit();
}
}
if (i == n)
{ //They both are equal;
}
else if(i != n)
{
//They both are not equal;
}
If it is not working please comment on it.
Thank You.

Here is a fix for the code that you posted.
for(int i = 0; i < n; i++)
{
int j;
for(j = 0; j < n; j++)
{
if(a[i] == b[j]) break;
}
if (j == n) return false;
}
return true;
That algorithm is going to break done for arrays that contain duplicates, however. For example, the array {1, 1, 2, 3} will be found as a match to the array {1, 2, 2, 3}.
I would highly recommend that you implement a sort-and-compare algorithm instead.