Cleaning and transformation

Normalise dates in Excel

Upload your sheet with dates in different formats and download it with all of them unified and recognised as real dates.

Free, no usage limits Nothing written to disk .xlsx,.xls,.csv

Accepted formats: .xlsx,.xls,.csv

Leave it empty to scan the whole sheet.

“03/04/2026” means two different dates depending on the order. Choose wrong and the file comes out with day and month swapped while looking perfectly fine: check the warning afterwards.

YYYY-MM-DD is the one every tool understands and the only one that sorts correctly as text.

The problem

When data comes from several places, the same column ends up like this:

12/03/2026
2026-03-12
12-Mar-26
03/12/2026

Excel recognises some as dates and others as text, so the column cannot be sorted, filtered by range, or used in a pivot table by month.

The day-versus-month trap

03/04/2026 means two different dates: 3 April if the day comes first, 4 March if the month does. There is no way to tell by looking at the value; it depends on who produced the file.

That is why you have to tell the tool how they are written NOW. Choose wrong and the file comes out with day and month swapped and looks perfectly correct. It is the most expensive mistake in this category.

To protect you, the tool counts how many dates were ambiguous (day and month both below 13) and tells you afterwards. If that number is high, check a couple of rows before trusting the file.

Why YYYY-MM-DD

It is the ISO 8601 format. Every database understands it, it holds no possible ambiguity, and it is the only one that sorts correctly even when treated as text.

Frequently asked questions

Do the dates come out as dates or as text?

As real dates, so Excel can sort and filter them. There is an ISO text output option if you need to export to a system expecting a string.

What about values that are not dates?

They are left untouched. Emptying a cell you do not understand would be destroying data.

Does it recognise month names?

Yes, in your language: "12-Mar-2026" is recognised without configuring anything.

What about dates that were already dates?

The chosen output format is applied without re-interpreting them: there is no ambiguity to resolve.