Question

How can I split a fullname into single, medium and last name?

Obviously, it is impossible to cover all posibilities. I only want one rule: If after the word there is other word from 3 letter or less, it must join it with the next word.

Also, I am supposing that the fullname has 3 words or more.

I really have no idea even how to start.

For example:

John Martin Jackson:

• name1: John
• name2: Martin
• name3: Jackson

Steven Ponce de Leon Presley

• name1: Steven
• name2: Ponce de Leon
• name3: Presley

Michael de la Rosa Martin Jackson:

• name1: Michael de la Rosa
• name2: Martin
• name3: Jackson

:S

Answer 1

A really fancy regex could do that. To match one name, use

/\S+((\s+\S{1,3})+\s+\S+)*/

Then, combine three of them with non-matching groups, but each wrapped in one, joined by whitespaces:

/^(\S+(?:(?:\s+\S{1,3})+\s+\S+)*)\s+(\S+(?:(?:\s+\S{1,3})+\s+\S+)*)\s+(\S+(?:(?:\s+\S{1,3})+\s+\S+)*)$/

To make it match people without a middle name, make that optional:

/^(\S+(?:(?:\s+\S{1,3})+\s+\S+)*)(?:\s+(\S+(?:(?:\s+\S{1,3})+\s+\S+)*))?\s+(\S+(?:(?:\s+\S{1,3})+\s+\S+)*)$/

Update: Don’t try to match whole names in one regex. Just use the first (simple) regex with a global flag:

> "Steven Ponce de Leon Presley".match(/\S+((\s+\S{1,3})+\s+\S+)*/g)
["Steven", "Ponce de Leon", "Presley"]

Explanation:

/
 \S+      match a word
 (         followed by any number of
  (         at least one
   \s+       whitespace-separated
   \S{1,3}   up-to-three-letters word
  )+
  \s+       and a whitespace-separated
  \S+       word
 )*
/g

---

However, I think an algorithm with some string and array functions would make it clearer what happens, and allows more customisation of the matching process:

var names = input.split(/s+/);
if (names.length < 2)
    return; // special case handling for only one word
var short = 0;
for (var i=names.length-2; i>=0; i--) {
    // starting at the second-to-last, I expect names not to end with a short one
    if (names[i].length < 4) {
        short++;
    } else if (short) {
        names[i] += " "+names.splice(i+1, short+1).join(" ");
        short = 0;
    }
}
return names; // an Array with at least one name