Replace all occurrences of a string in Javascript + A non-breaking solution

Say you have a string. You want to remove/replace all instances of a character in that string. You could use the replace method in Javascript.

It looks something like this:

String.replace(/<TERM>/g, '');

For example,

'Hello-World'.replace(/-/g, ' ');
// "Hello World"

'H E L L O'.replace(/ /g, '');
// "HELLO"

Note that this does a case-sensitive substitution. For a case-insensitive substitution, use the /i flag. For example:

'AbBle'.replace(/b/gi, 'p');
// "Apple"

This is a super useful method and a little understanding of regular expressions will help you write more concise, performant code.

But, what if the string is null or undefined? I recently had a use-case where the string was a variable that could be undefined or null. To be precise,

object['property'].replace(/ /g, '');

And, object["property"] came from an API. So, the returned object either has the property or doesn't.

If object["property"] === null or object["property"] === undefined, you see an ugly TypeError:

Using .replace on undefined or null

In this case, I don't want the application to throw an error if object["property"] === null or object["property"] === undefined. Still, if object["property"] is a valid string, I wanted to modify it.

So, I wrote my own replace function:

const replaceString = (str, newSubstring, regex) => {
  if (!str) {
    return str;
  return str.replace(regex, newSubstring);

replaceString(undefined, '', /-/g);
// undefined

replaceString(null, '', /-/g);
// null

replaceString('H-E-L-L-O', '', /-/g);
// "HELLO"

I could have redefined the String prototype method for my module instead. However, I wanted to avoid confusions in the future when a coder wants to use the .replace function as originally defined.