Employee Name Splitter
One name column into title, first, middle, last and suffix, with the rows it will not vouch for shown first rather than buried.
Reads .xlsx and .csv inside your browser. Nothing is uploaded.
This cannot be done correctly, and a tool that pretends otherwise does real harm
Getting a name wrong is not a formatting error. It is somebody’s name entered incorrectly in the system that pays them, prints their badge and addresses their post, and it stays that way for years because nobody re-checks a field that filled itself in. So the most important number on this page is how many rows need a person to look at them.
The name list
Drop a spreadsheet here
One row per person, with their name in a single column.
Cutting at the last space is wrong more often than you think
Ludwig van Beethoven’s surname is “van Beethoven”, not “Beethoven”. Pieter van der Berg’s is three words. Maria de la Cruz’s is three words. Ahmed bin Rashid’s is two. Every one of those comes out wrong from the obvious approach, and every one of them is a common name.
Words like van, von, de, del, bin and mac begin a surname rather than standing alone, and they are handled here. Words like the Spanish “y” do a different job: they join two surnames, so the surname starts one word earlier than the connector rather than at it. Treating those two cases the same way loses a surname.
Some of it is genuinely undecidable
“Maria Teresa Garcia Lopez” might be a first name, a middle name and a double surname, or two given names and two surnames, or several other things. The string does not say. “Li Wei” puts the family name first, and so do a great many names, and nothing inside the text distinguishes them from a name written the other way round. A mononym is a complete name, not half of one.
Those rows get a best guess and a plain statement of why it is only a guess. They are listed above the rows that worked, because a list of exceptions underneath four hundred successes is a list nobody reads.
Short particles have to be lowercase
“al” is a particle in Hassan al Rashid and is not one in Al Smith. “Y” is a connector in Ana Garcia y Lopez and a middle initial in John Y Smith. Two-letter particles are therefore only treated as particles when the original is lowercase, which is the convention in every case where they really are one.
Nothing is changed, only separated
Capitalisation is left exactly as it arrived, because deciding that MACDONALD should become MacDonald rather than Macdonald is a guess about one particular family. The original string is kept beside every result, so nothing is ever lost and any row can be checked against what was actually in the file.