Develop Reference

Sum of similar value in n X n dimensional array with n^2 complexity

Given an array [[1, 7, 3, 8],[3, 2, 9, 4],[4, 3, 2, 1]],
how can I find the sum of its repeating elements? (In this case, the sum would be 10.)
Repeated values are - 1 two times, 3 three times, 2 two times, and 4 two times
So, 1 + 3 + 2 + 4 = 10
Need to solve this problem in the minimum time
There are multiple ways to solve this but time complexity is a major issue.
I try this with the recursion function
How can I optimize more
`
var uniqueArray = []
var sumArray = []
var sum = 0
function sumOfUniqueValue (num){
for(let i in num){
if(Array.isArray(num[i])){
sumOfUniqueValue(num[i])
}
else{
// if the first time any value will be there then push in a unique array
if(!uniqueArray.includes(num[i])){
uniqueArray.push(num[i])
}
// if the value repeats then check else condition
else{
// we will check that it is already added in sum or not
// so for record we will push the added value in sumArray so that it will added in sum only single time in case of the value repeat more then 2 times
if(!sumArray.includes(num[i])){
sumArray.push(num[i])
sum+=Number(num[i])
}
}
}
}
}
sumOfUniqueValue([[1, 7, 3, 8],[1, 2, 9, 4],[4, 3, 2, 7]])
console.log("Sum =",sum)
`
That's a real problem, I am just curious to solve this problem so that I can implement it in my project.
If you guys please mention the time it will take to complete in ms or ns then that would be really helpful, also how the solution will perform on big data set.
Thanks

I would probably use a hash table instead of an array search with .includes(x) instead...
And it's also possible to use a classical for loop instead of recursive to reduce call stack.
function sumOfUniqueValue2 (matrix) {
const matrixes = [matrix]
let sum = 0
let hashTable = {}
for (let i = 0; i < matrixes.length; i++) {
let matrix = matrixes[i]
for (let j = 0; j < matrix.length; j++) {
let x = matrix[j]
if (Array.isArray(x)) {
matrixes.push(x)
} else {
if (hashTable[x]) continue;
if (hashTable[x] === undefined) {
hashTable[x] = false;
continue;
}
hashTable[x] = true;
sum += x;
}
}
}
return sum
}
const sum = sumOfUniqueValue2([[1, 7, 3, 8],[[[[[3, 2, 9, 4]]]]],[[4, 3, 2, 1]]]) // 10
console.log("Sum =", sum)
This is probably the fastest way...
But if i could choose a more cleaner solution that is easier to understand then i would have used flat + sort first, chances are that the built in javascript engine can optimize this routes instead of running in the javascript main thread.
function sumOfUniqueValue (matrix) {
const numbers = matrix.flat(Infinity).sort()
const len = numbers.length
let sum = 0
for (let i = 1; i < len; i++) {
if (numbers[i] === numbers[i - 1]) {
sum += numbers[i]
for (i++; i < len && numbers[i] === numbers[i - 1]; i++);
}
}
return sum
}
const sum = sumOfUniqueValue2([[1, 7, 3, 8],[[[[[3, 2, 9, 4]]]]],[[4, 3, 2, 1]]]) // 10
console.log("Sum =", sum)

You could use an objkect for keeping trak of seen values, like
seen[value] = undefined // value is not seen before
seen[value] = false // value is not counted/seen once
seen[value] = true // value is counted/seen more than once
For getting a value, you could take two nested loops and visit every value.
Finally return sum.
const
sumOfUniqueValue = (values, seen = {}) => {
let sum = 0;
for (const value of values) {
if (Array.isArray(value)) {
sum += sumOfUniqueValue(value, seen);
continue;
}
if (seen[value]) continue;
if (seen[value] === undefined) {
seen[value] = false;
continue;
}
seen[value] = true;
sum += value;
}
return sum;
},
sum = sumOfUniqueValue([[1, 7, 3, 8], [3, 2, 9, 4], [4, 3, 2, 1]]);
console.log(sum);
Alternatively take a filter and sum the values. (it could be more performat with omitting same calls.)
const
data = [[1, 7, 3, 8], [3, 2, 9, 4, 2], [4, 3, 2, 1]],
sum = data
.flat(Infinity)
.filter((v, i, a) => a.indexOf(v) !== a.lastIndexOf(v) && i === a.indexOf(v))
.reduce((a, b) => a + b, 0);
console.log(sum);

You can flatten the array, filter-out single-instance values, and sum the result:
const data = [
[ 1, 7, 3, 8 ],
[ 3, 2, 9, 4 ],
[ 4, 3, 2, 1 ]
];
const numbers = new Set( data.flat(Infinity).filter(
(value, index, arr) => arr.lastIndexOf(value) != index)
);
const sum = [ ...numbers ].reduce((a, b) => a + b, 0);
Another approach could be the check the first and last index of the number in a flattened array, deciding whether or not it ought to be added to the overall sum:
let sum = 0;
const numbers = data.flat(Infinity);
for ( let i = 0; i < numbers.length; i++ ) {
const first = numbers.indexOf( numbers[ i ] );
const last = numbers.lastIndexOf( numbers[ i ] );
if ( i == first && i != last ) {
sum = sum + numbers[ i ];
}
}
// Sum of numbers in set
console.log( sum );

What is the efficient way to find an array inside an array?

Find an Array of Values Inside an Array
Lets say I have an array [1,2,3,8,2,3,4,5,6,7,8] and I want to find the first occurrence of the values [3,4,5,6] together, how might I do that? I can use Array.prototype.findIndex, but when I am looking for a large amount of values in a large array, it doesn't feel like the proper way to do it.
What fails:
let largeArray = [1,2,3,8,2,3,4,5,6,7,8];
let smallArray = [3,4,5,6];
//Problem: an array isn't a function
largeArray.findIndex(smallArray);
/*
Problem: always returns -1 because it checks each element
rather than looking for a group of elements.
*/
largeArray.indexOf(smallArray);
//Problem: same as using indexOf
largeArray.findIndex(item=>smallArray);
What works:
let largeArray = [1,2,3,8,2,3,4,5,6,7,8];
let smallArray = [3,4,5,6];
//Here is what works, but isn't ideal
largeArray.findIndex((item, index, arr) => {
let isTheOne = item == smallArray[0] &&
arr[index + 1] == smallArray[1] &&
arr[index + 2] == smallArray[2] &&
arr[index + 3] == smallArray[3];
return isTheOne;
});
//It returns 5, which is correct.
To Be Continued
I am currently using what works, but what if largeArray had the length of a million and smallArray had the length of 300. That would be 1 line of item == smallArray[0] &&, 298 lines of arr[index + x] == smallArray[x] &&, and 1 line of arr[index + x] == smallArray[x];. I don't want to use Array.prototype.map, Array.prototype.filter, Array.prototype.forEach, a for loop, or a while loop. This is because Array.prototype.map, Array.prototype.forEach, and the loops take a very long time to complete. I don't want to use Array.prototype.filter because that doesn't give me the index.

You were on the right track, you just want to use every() to look over the small index to check that each index matches
const largeArray = [1, 2, 3, 8, 2, 3, 4, 5, 6, 7, 8];
let smallArray = [3, 4, 5, 6];
const index = largeArray.findIndex(
(item, index, arr) =>
smallArray.every(
(n, sIndex) => n === arr[index + sIndex]
)
);
console.log(index);
You could add a check beforehand to not have to go in every... not sure what that would improve.
const index = largeArray.findIndex(
(item, index, arr) =>
item === smallArray[0] &&
smallArray.every(
(n, sIndex) => n === arr[index + sIndex]
)
);
Other approach is using strings
const largeArray = [1, 2, 3, 8, 2, 3, 4, 5, 6, 7, 8];
const smallArray = [3, 4, 5, 6];
const largeStr = largeArray.join(",");
const smallStr = smallArray.join(",");
const strIndex = largeStr.indexOf(smallStr);
const index = strIndex > -1 ? largeStr.substr(0,strIndex-1).split(",").length : -1;
console.log(index)
To figure out what is better is really based on your use case.

You can use .join to convert the arrays to strings, and use .indexOf to get the index given that you will remove the additional commas:
const getIndexOfSubArray = (arr=[], sub=[]) => {
const str = arr.join();
const subStr = sub.join();
const index = str.indexOf(subStr);
return index < 0 ? -1 : str.substr(0, index-1).split(',').length;
}
console.log( getIndexOfSubArray([1,2,3,8,2,3,4,5,6,7,8], [3,4,5,6]) );

You could iterate by hand and check the items with indexOf.
function getIndex(array, subarray) {
let p = -1,
first = subarray[0];
while ((p = array.indexOf(first, p + 1)) !== -1) {
let i = p,
complete = true;
for (const s of subarray) {
if (s !== array[i++]) {
complete = false;
break;
}
}
if (complete) return p;
}
return -1;
}
console.log(getIndex([1, 2, 3, 8, 2, 3, 4, 5, 6, 7, 8], [3, 4, 5, 6])); // 5

Here is a simple approach to this problem:
let largeArray = [1, 2, 3, 8, 2, 3, 4, 5, 6, 7, 8];
let smallArray = [3, 4, 5, 6];
let s = 0,
i = 0,
j = 0;
let SLen = smallArray.length,
LLen = largeArray.length;
while (i < LLen && j < SLen && SLen - j <= LLen - i) {
if (j == 0) {
s = i;
}
if (largeArray[i] == smallArray[j]) {
j++;
} else {
j = 0;
i = s;
}
i++;
}
let index = i - j;
if (j == SLen) {
console.log(`found at index ${index}`);
} else {
console.log('not found');
}

How to write the code with less time complexity for finding the missing element in given array range?

My function should return the missing element in a given array range.
So i first sorted the array and checked if the difference between i and i+1 is not equal to 1, i'm returning the missing element.
// Given an array A such that:
// A[0] = 2
// A[1] = 3
// A[2] = 1
// A[3] = 5
// the function should return 4, as it is the missing element.
function solution(A) {
A.sort((a,b) => {
return b<a;
})
var len = A.length;
var missing;
for( var i = 0; i< len; i++){
if( A[i+1] - A[i] >1){
missing = A[i]+1;
}
}
return missing;
}
I did like above, but how to write it more efficiently??

You could use a single loop approach by using a set for missing values.
In the loop, delete each number from the missing set.
If a new minimum value is found, all numbers who are missing are added to the set of missing numbers, except the minimum, as well as for a new maximum numbers.
The missing numbers set contains at the end the result.
function getMissing(array) {
var min = array[0],
max = array[0],
missing = new Set;
array.forEach(v => {
if (missing.delete(v)) return; // if value found for delete return
if (v < min) while (v < --min) missing.add(min); // add missing min values
if (v > max) while (v > ++max) missing.add(max); // add missing max values
});
return missing.values().next().value; // take the first missing value
}
console.log(getMissing([2, 3, 1, 5]));
console.log(getMissing([2, 3, 1, 5, 4, 6, 7, 9, 10]));
console.log(getMissing([3, 4, 5, 6, 8]));

Well, from the question (as it's supposed to return a single number) and all the existing solution (examples at least), it looks like list is unique. For that case I think we can sumthe entire array and then subtracting with the expected sum between those numbers will generate the output.
sum of the N natural numbers
1 + 2 + ....... + i + ... + n we can evaluate by n * (n+1) / 2
now assume, in our array min is i and max is n
so to evaluate i + (i+1) + ..... + n we can
A = 1 + 2 + ..... + (i-1) + i + (i+1) + .... n (i.e. n*(n+1)/2)
B = 1 + 2 + ..... + (i-1)
and
C = A - B will give us the sum of (i + (i+1) + ... + n)
Now, we can iterate the array once and evaluate the actual sum (assume D), and C - D will give us the missing number.
Let's create the same with each step at first (not optimal for performance, but more readable) then we will try to do in a single iteration
let input1 = [2, 3, 1, 5],
input2 = [2, 3, 1, 5, 4, 6, 7, 9, 10],
input3 = [3, 4, 5, 6, 8];
let sumNatural = n => n * (n + 1) / 2;
function findMissing(array) {
let min = Math.min(...array),
max = Math.max(...array),
sum = array.reduce((a,b) => a+b),
expectedSum = sumNatural(max) - sumNatural(min - 1);
return expectedSum - sum;
}
console.log('Missing in Input1: ', findMissing(input1));
console.log('Missing in Input2: ', findMissing(input2));
console.log('Missing in Input3: ', findMissing(input3));
Now, lets try doing all in a single iteration (as we were iterating 3 times for max, min and sum)
let input1 = [2, 3, 1, 5],
input2 = [2, 3, 1, 5, 4, 6, 7, 9, 10],
input3 = [3, 4, 5, 6, 8];
let sumNatural = n => n * (n + 1) / 2;
function findMissing(array) {
let min = array[0],
max = min,
sum = min,
expectedSum;
// assuming the array length will be minimum 2
// in order to have a missing number
for(let idx = 1;idx < array.length; idx++) {
let each = array[idx];
min = Math.min(each, min); // or each < min ? each : min;
max = Math.max(each, max); // or each > max ? each : max;
sum+=each;
}
expectedSum = sumNatural(max) - sumNatural(min - 1);
return expectedSum - sum;
}
console.log('Missing in Input1: ', findMissing(input1));
console.log('Missing in Input2: ', findMissing(input2));
console.log('Missing in Input3: ', findMissing(input3));

Instead of sorting, you could put each value into a Set, find the minimum, and then iterate starting from the minimum, checking if the set has the number in question, O(N). (Sets have guaranteed O(1) lookup time)
const input1 = [2, 3, 1, 5];
const input2 = [2, 3, 1, 5, 4, 6, 7, 9, 10];
const input3 = [3, 4, 5, 6, 8];
function findMissing(arr) {
const min = Math.min(...arr);
const set = new Set(arr);
return Array.from(
{ length: set.size },
(_, i) => i + min
).find(numToFind => !set.has(numToFind));
}
console.log(findMissing(input1));
console.log(findMissing(input2));
console.log(findMissing(input3));

If the array is items and the difference between missing and present diff is 1:
const missingItem = items => [Math.min(...items)].map(min => items.filter(x =>
items.indexOf(x-diff) === -1 && x !== min)[0]-diff)[0]
would give complexity of O(n^2).
It translates to: find the minimum value and check if there isn't a n-diff value member for every value n in the array, which is also not the minimum value. It should be true for any missing items of size diff.
To find more than 1 missing element:
([Math.min(...items)].map(min => items.filter(x =>
items.indexOf(x-diff) === -1 && x !== min))[0]).map(x => x-diff)
would give O((m^2)(n^2)) where m is the number of missing members.

Found this old question and wanted to take a stab at it. I had a similar thought to https://stackoverflow.com/users/2398764/koushik-chatterjee in that I think you can optimize this by knowing that there's always only going to be one missing element. Using similar methodology but not using a max will result in this:
function getMissing(arr) {
var sum = arr.reduce((a, b) => a + b, 0);
var lowest = Math.min(...arr);
var realSum = (arr.length) * (arr.length + 1) / 2 + lowest * arr.length;
return realSum - sum + lowest;
}
With the same inputs as above. I ran it in jsperf on a few browsers and it is faster then the other answers.
https://jsperf.com/do-calculation-instead-of-adding-or-removing.
First sum everything, then calculate the lowest and calculate what the sum would be for integers if that happened to be the lowest. So for instance if we have 2,3,4,5 and want to sum them that's the same as summing 0,1,2,3 and then adding the lowest number + the amount of numbers in this case 2 * 4 since (0+2),(1+2),(2+2),(3+2) turns it back into the original. After that we can calculate the difference but then have to increase it once again by the lowest. To offset the shift we did.

You can use while loop as well, like below -
function getMissing(array) {
var tempMin = Math.min(...array);
var tempMax = tempMin + array.length;
var missingNumber = 0;
while(tempMin <= tempMax){
if(array.indexOf(tempMin) === -1) {
missingNumber = tempMin;
}
tempMin++;
}
return missingNumber;
}
console.log(getMissing([2, 3, 1, 5]));
console.log(getMissing([2, 3, 1, 5, 4, 6, 7, 9, 10]));
console.log(getMissing([3, 4, 5, 6, 8]));

My approach is based on in place sorting of the array which is O(N) and without using any other data structure.
Find the min element in the array.
Sort in place.
Again loop the array and check if any element is misplaced, that is the answer!
function getMissing(ar) {
var mn = Math.min(...ar);
var size = ar.length;
for(let i=0;i<size;i++){
let cur = ar[i];
// this ensures each element is in its right place
while(cur != mn +i && (cur - mn < size) && cur != ar[cur- mn]) {
// swap
var tmp = cur;
ar[i] = ar[cur-mn];
ar[cur-mn] = tmp;
}
}
for(let i=0; i<size; i++) {
if(ar[i] != i+mn) return i+mn;
}
}
console.log(getMissing([2, 3, 1, 5]));
console.log(getMissing([2, 3, 1, 5, 4, 6, 7, 9, 10]));
console.log(getMissing([3, 4, 5, 6, 8]));

JavaScript: Writing this solution using higher order functions

I worked on a problem where you are given an array of numbers and a target sum, and it's your job to find a pair of numbers that sum up to the target number. Here was my solution using simple nested for loops:
function findPairForSum(integers, target) {
var output = [];
for (var i = 0; i < integers.length; i++) {
for (var j = 0; j < integers.length; j++) {
if (i !== j && integers[i] + integers[j] === target) {
output.push(integers[i], integers[j]);
return output;
}
}
}
return 'not possible';
}
findPairForSum([3, 34, 4, 12, 5, 2], 9); // --> [4, 5]
My question is, is there a cleaner way to write this solution using higher order functions (perhaps forEach?)
Here was my attempt to use forEach:
function findPairForSum(integers, target) {
var output = [];
integers.forEach(function(firstNum) {
integers.forEach(function(secondNum) {
if (firstNum + secondNum === target) {
output.push(firstNum, secondNum);
}
})
})
if (output === []) {
return 'not possible';
}
return output;
}
findPairForSum([3, 34, 4, 12, 5, 2], 9); // --> [ 4, 5, 5, 4 ]
I tried putting a return after the two pushes, but it did not return anything. So instead, I put the return at the very end.
Why won't it return after the initial two pushes? I want it to stop right there, and only push the two numbers. Instead, by putting the return at the end, it pushed 4 numbers. It should be [4,5] but I got something like [4,5,5,4].
Any advice and help would be much appreciated!

Assume we have the following set of numbers, and we must find a subset of 2 numbers whose sum is 9:
Numbers: 4, 5, 6
Your current code iterates both with i and j from 0 to length. This means that the following iterations match the condition:
Indices: 0, 1, 2
Numbers: 4, 5, 6 // (i) (j)
---------------- // ↓ ↓
i j // Numbers[0] + Numbers[1] === 9
j i // Numbers[1] + Numbers[0] === 9
As you can see, the numbers 4 and 5 are matched twice, in 2 iterations:
i === 0 && j === 1
i === 1 && j === 0
You can avoid this by making sure one simple condition is met:
j must at all times be greater than i
This condition can be met met by initializing j with i + 1 in the inner for loop:
for (var i = 0; i < integers.length; i++) {
for (var j = i + 1; j < integers.length; j++) {
// ...
}
}
This way, j can never be 0 when i is 1, because the inner for-loop will run to completion before i is ever incremented once more. Once that happens, a brand new inner for-loop is created, in which j is again set to i + 1. The following diagram is the result:
Indices: 0, 1, 2
Numbers: 4, 5, 6
----------------
i j
X i // ← j can never be 0 if (i === 1),
// so the same set is never evaluated twice.
In other words, only the following combinations for i and j are checked at most:
Indices: 0, 1, 2
----------------
i j
i j
i j
is there a cleaner way to write this solution using higher order functions (perhaps forEach?)
A for loop is actually a fine solution for your use-case. They allow you to break early - after the first time you find a valid pair of numbers. forEach or other array iterator functions on the other hand will always continue until all set indices are visited.
You are actually breaking early in your first example with the statement return output;
When you use forEach on a set of numbers with multiple valid sets, you'll always get back all numbers involved:
Indices: 0, 1, 2, 3
Numbers: 4, 5, 6, 3 // (i) (j)
------------------- // ↓ ↓
i j // Numbers[0] + Numbers[1] === 4 + 5 === 9
i j // Numbers[2] + Numbers[3] === 6 + 3 === 9
forEach, map, reduce and the like do not allow you to break early. The following snippet demonstrates this issue of the diagram above:
function findPairForSum(integers, target) {
var output = [];
integers.forEach(function(firstNum, i) {
// slice(i + 1) has the same effect as for (var j = i + 1; ...)
integers.slice(i + 1).forEach(function(secondNum, j) {
if (firstNum + secondNum === target) {
// There is no way here to stop the iteration of either
// forEach call... T_T
output.push(firstNum, secondNum);
}
});
})
if (output.length) {
return output;
}
return 'not possible';
}
console.log(findPairForSum([4, 5, 6, 3], 9)); // --> [4, 5, 6, 3]
This is why I highly recommend sticking with the for loops for this specific use case. With for loop you can simply return as you already did as soon as you encounter a valid set of numbers:
function findPairForSum(integers, target) {
for (var i = 0; i < integers.length; i++) {
for (var j = i + 1; j < integers.length; j++) {
if (integers[i] + integers[j] === target) {
return [integers[i], integers[j]];
}
}
}
return 'not possible';
}
console.log(findPairForSum([4, 5, 6, 3], 9)); // --> [4, 5]

This could be your solution:
function findPairForSum(arr, sum) {
var pairs = [];
arr.forEach(n1 => {
var n2 = arr.find(n2 => n1 + n2 == sum)
if (n2) pairs.push([n1, n2]);
});
return pairs;
}
var sums = findPairForSum([3, 34, 4, 12, 6, 2], 9);
console.log(sums)

The problem is, you iterate from the start of the array for the inner loop. You could use a copy which starts at the index of the outer loop plus one and exit early on a found value.
But this does not solves the problem with multiple pairs. The result is simply wrong.
function findPairForSum(integers, target) {
var output = [];
integers.forEach(function(firstNum, i) {
integers.slice(i + 1).some(function(secondNum) {
if (firstNum + secondNum === target) {
output.push(firstNum, secondNum);
return true;
}
});
});
return output.length && output || 'not possible';
}
// console.log(findPairForSum([3, 34, 4, 12, 5, 2], 9));
console.log(findPairForSum([3, 34, 4, 4, 12, 5, 2, 4, 5], 9));
For a solution, you need to remember which pairs are used. This approach works with only one loop and a hash table for counting missing values.
If a pair is found, the counter is decremented and the two values are pushed to the result set.
function findPairForSum(integers, target) {
var hash = Object.create(null),
output = [];
integers.forEach(function(value) {
if (hash[value]) {
output.push(target - value, value);
hash[value]--;
return;
}
hash[target - value] = (hash[target - value] || 0) + 1;
});
return output.length && output || 'not possible';
}
console.log(findPairForSum([3, 34, 4, 4, 12, 5, 2, 4, 5], 9));

This is expected, since you didn't compare the indexes.
This inner array should only loop through the indexes which larger than the outer index.
You can achieve this by using the 2nd parameter, index, in forEach's callback function:
const ints = [3, 34, 4, 12, 5, 6, 2];
function findPairForSum(integers, target) {
let result;
integers.forEach((val1, idx1) => {
integers.forEach((val2, idx2) => {
if (idx1 < idx2 && val1 + val2 === target) {
result = [val1, val2];
}
})
})
return result;
}
console.log(findPairForSum(ints, 9));

Use can reduce your array into another which has sum equals target value:
const ints = [3, 34, 4, 12, 6, 2];
const value = 9;
const resp = ints.reduce((acc, ele, idx, self) => {
let found = self.find(x => x + ele == value)
return found ? [found, ele] : acc;
}, []);
console.log(resp); // [3, 6]

You can use Array.prototype.some which will stop execution as soon as the condition becomes true. See below code.
function findPairForSum(arr, sum) {
var pairs = [];
arr.some(n1 => {
var n2 = arr.find(n2 => n1 + n2 == sum)
if (n2) {
pairs.push(n1, n2); return true;
};
return false;
});
return pairs.length > 0 ? pairs : "not possible";
}
console.log(findPairForSum([3, 34, 4, 12, 7, 2], 9));

JS: Replace null value in array with an average of the previous and next value

If I have the following array
[5,2,null,5,9,4]
I need to replace null with the average of the previous and next values (2 and 5). Like so:
[5,2,3.5,5,9,4]
Additionally, if I had an array with consecutive null values:
[5,2,null,null,9,4]
I would like to be able to replace them with the step values between the 2 and 9. Like so:
[5,2,4.33,6.63,9,4]
Edit: Guys, totally not a homework assignment. I'm trying to do some stuff with the Chartist-js library.
Here's what I've got that works to replace null with the values before and after.
var seriesData = [5,2,null,5,null,4]
seriesData.forEach(function(value, index){
if ( value === null ) {
var prev = seriesData[index - 1];
var next = seriesData[index + 1];
seriesData[index] = (prev + next) / 2;
}
});
=> [5, 2, 3.5, 5, 4.5, 4]

This does sound like a Chart problem, so here's a crude logic, you can easily improve the code:
function test(){
var arr = [5,3,2,null,null,9,4, null, null, null, 10, null, 12],
ind = -1,
prevIndex,
nJumps, valJump;
for( var currIndex = 0; currIndex< arr.length; currIndex++ ){
if( arr[ currIndex] == null ){
continue;
}else if( prevIndex === undefined || (nJumps = (currIndex- prevIndex)) === 1 ){
prevIndex = currIndex;
continue;
}
valJump = (( arr[ currIndex ] - arr[ prevIndex ]) / nJumps);
ind = (+prevIndex);
while( ++ind < (+currIndex) )
arr[ ind ] = (arr[ ind-1 ]) + valJump;
prevIndex = currIndex;
}
return arr;
}

This solves your first problem:
var arr = [5, 2, null, 5, 9, 4];
var result = arr
.map(function (x) {
if (x == null) {
return (arr[arr.indexOf(x) - 1] + arr[arr.indexOf(x) + 1]) / 2 ;
}
return x;
});
console.log(result); // [5, 2, 3.5, 5, 9, 4]
For the second problem I don't quite understand what you mean with the two steps stuff.