LeetCode Meditations: Palindromic Substrings

The description for Palindromic Substrings is:

Given a string s, return the number of palindromic substrings in it.

A string is a palindrome when it reads the same backward as forward.

A substring is a contiguous sequence of characters within the string.

For example:

Input: s = "abc"
Output: 3
Explanation: Three palindromic strings: "a", "b", "c".

Or:

Input: s = "aaa"
Output: 6
Explanation: Six palindromic strings: "a", "a", "a", "aa", "aa", "aaa".

Also, the constraints indicate that s consists of lowercase English letters.

---

In the previous problem, we found a solution to find the longest palindromic substring in a given string. To find a palindrome, we used an "expand over center" approach, where we assumed each character as the middle character of a substring. And, accordingly, we shifted the left and right pointers.

Note
Checking a palindrome is easy with two pointers approach, which we've seen before with Valid Palindrome.

Counting the palindromes in one substring may look like this:

function countPalindromesInSubstr(s: string, left: number, right: number): number {
  let result = 0;
  while (left >= 0 && right < s.length && s[left] === s[right]) {
    result++;
    left--;
    right++;
  }

  return result;
}

As long as we're within bounds (left >= 0 && right < s.length), we check if two characters on the left and right are the same — if so, we update our result and shift our pointers.

However, once you think about it, it matters which indices the pointers are initialized at. For example, if we pass the string "abc" to countPalindromesInSubstr, and the left pointer is at 0 while the right pointer is at the last index (2), then our result is simply 0.

Remember that we are assuming each character as the middle character of a substring, and since each single character is also a substring as well, we'll initialize our left and right pointers to point to the character itself.

Note
A character by itself is considered palindromic, i.e., "abc" has three palindromic substrings: 'a', 'b' and 'c'.

---

Let's see how this process might look like.

If we have the string "abc", we'll first assume that 'a' is the middle of a substring:

"abc"

left = 0
right = 0
currentSubstr = 'a'

totalPalindromes = 1 // A single character is a palindrome

Then, we'll try to expand the substring to see if 'a' can be the middle character of another substring:

"abc"

left = -1
right = 1
currentSubstr = undefined

totalPalindromes = 1

Now that our left pointer is out of bounds, we can jump to the next character:

"abc"

left = 1
right = 1
currentSubstr = 'b'

totalPalindromes = 2

Now, we'll update our pointers, and indeed, 'b' can be the middle character of another substring:

s = "abc"

left = 0
right = 2
currentSubstr = 'abc'

totalPalindromes = 2

Well, currentSubstr is not a palindrome. Now we update our pointers again:

s = "abc"

left = -1
right = 3
currentSubstr = undefined

totalPalindromes = 2

And, we're out of bounds again. Time to move on to the next character:

s = "abc"

left = 2
right = 2
currentSubstr = 'c'

totalPalindromes = 3

Shifting our pointers, we'll be out of bounds again:

s = "abc"

left = 1
right = 3
currentSubstr = undefined

totalPalindromes = 3

Now that we've gone through each character, our final result of totalPalindromes is, in this case, 3. Meaning that there are 3 palindromic substrings in "abc".

However, there is an important caveat: each time we assume a character as the middle and initialize two pointers to the left and right of it, we're trying to find only odd-length palindromes. In order to mitigate that, instead of considering a single character as the middle, we can consider two characters as the middle and expand as we did before.

In this case, the process of finding the even-length substring palindromes will look like this — initially, our right pointer is left + 1:

s = "abc"

left = 0
right = 1
currentSubstr = 'ab'

totalPalindromes = 0

Then, we'll update our pointers:

s = "abc"

left = -1
right = 2
currentSubstr = undefined

totalPalindromes = 0

Out of bounds. On to the next character:

s = "abc"

left = 1
right = 2
currentSubstr = 'bc'

totalPalindromes = 0

Updating our pointers:

s = "abc"

left = 0
right = 3
currentSubstr = undefined

totalPalindromes = 0

The right pointer is out of bounds, so we go on to the next character:

s = "abc"

left = 2
right = 3
currentSubstr = undefined

totalPalindromes = 0

Once again we're out of bounds, and we're done going through each character. There are no palindromes for even-length substrings in this example.

---

We can write a function that does the work of counting the palindromes in each substring:

function countPalindromes(s: string, isOddLength: boolean): number {
  let result = 0;
  for (let i = 0; i < s.length; i++) {
    let left = i;
    let right = isOddLength ? i : i + 1;
    result += countPalindromesInSubstr(s, left, right);
  }

  return result;
}

In our main function, we can call countPalindromes twice for both odd and even length substrings, and return the result:

function countSubstrings(s: string): number {
  let result = 0;
  result += countPalindromes(s, true); // Odd-length palindromes
  result += countPalindromes(s, false); // Even-length palindromes

  return result;
}

Overall, our solution looks like this:

function countSubstrings(s: string): number {
  let result = 0;
  result += countPalindromes(s, true); // Odd-length palindromes
  result += countPalindromes(s, false); // Even-length palindromes

  return result;
}

function countPalindromes(s: string, isOddLength: boolean): number {
  let result = 0;
  for (let i = 0; i < s.length; i++) {
    let left = i;
    let right = isOddLength ? i : i + 1;
    result += countPalindromesInSubstr(s, left, right);
  }

  return result;
}

function countPalindromesInSubstr(s: string, left: number, right: number): number {
  let result = 0;
  while (left >= 0 && right < s.length && s[left] === s[right]) {
    result++;
    left--;
    right++;
  }

  return result;
}

Time and space complexity

The time complexity is $O(n^2)$ as we go through each substring for each character (countPalindromes is doing an $O(n^2)$ operation, we call it twice separately.)
The space complexity is $O(1)$ as we don't have an additional data structure whose size will grow with the input size.

---

Next up is the problem called Decode Ways. Until then, happy coding.