Coding Challenge: Return The Smaller String

Question

A coding challenge in which we are to write a function that compares two strings and returns the one that is smaller. The comparison is both lexicographical and numerical, depending on the content of the strings as explained below and in the comments of the code.

Both strings may contain any characters. Consecutive numbers in the string are considered a single number. If during the search the character of only one string is a number, then that string is returned as numbers are lexicographically smaller than letters.

If during the search the characters of both strings are numbers, then parseInt(strX.slice(i)) checks if there are more consecutive digits in each string and returns the string whose number is numerically smaller.

Examples:

input: "a", "b" expected output: "a" since "a" comes before "b" alphabetically

input: "a1", "a2" expected output: "a1" since 1 comes before 2

input: "a10", "a2" expected output: "a2" since 2 comes before 10

Here is the code:

const smallestString = (str1, str2) => {
    // we only need to iterate through the shortest string
    const len = str1.length < str2.length ? str1.length : str2.length;
    for (let i = 0; i < len; i++) {
        // check if both letters are strings
        if (str1[i].toUpperCase() !== str1[i].toLowerCase() && str2[i].toUpperCase() !== str2[i].toLowerCase()) {
            if (str1[i] === str2[i]) { // if both letters are the same, continue
                continue;
            } else {
                return str1[i] < str2[i] ? str1 : str2; // otherwise return the string with the 'smaller' char at that index
            }
        } else if (!isNaN(str1[i] || !isNaN(str2[i]))) { // check if either char is a number
            if (!isNaN(str1[i]) && !isNaN(str2[i])) { // if both chars are numbers, return str with numerically smaller number
              return parseInt(str1.slice(i)) < parseInt(str2.slice(i)) ? str1 : str2;
            }
        } else {
            return str1[i] < str2[i] ? str1 : str1;
        }
    }
    return str1; // if we get here then both strings are equal and either can be returned
}

console.log(smallestString('a10', 'a2')); // returns 'a2'

I am seeking any and all feedback about cleaning up the code and possibly improving the algorithm's current time complexity of O(n).

Answer 1

        if (!isNaN(str1[i]) && !isNaN(str2[i])) { // if both chars are numbers, return str with numerically smaller number
          return parseInt(str1.slice(i)) < parseInt(str2.slice(i)) ? str1 : str2;
        }

seems like a bug. In case the numbers compare equal the code blindly returns the first string. Consider (a10c, a10b).

---

return str1; // if we get here then both strings are equal and either can be returned

seems like another bug. At this point we only know that the strings are equal up to the length of the shortest one, and the shortest one should be returned. Which one is shortest is not tested here.

---

The overall logic looks overcomplicated. Consider testing for numbers first; that would eliminate the need for a confusing toUpper/toLower mess, and the special case of special characters.