Upload a spreadsheet with a date column that's a mix of formats — or a column of plain numbers that used to be dates — and get a new column with every value written the same way, so sorting and filtering actually work.
A date column built up from copy-pasted rows, several contributors, or a file that lost its formatting along the way is rarely one format. Sorting or filtering on it becomes unreliable in ways that are easy to miss until a report comes out wrong. This tool reads every value in the column you choose, works out what date it actually is — including the most common real cause, an Excel serial number like 45382 sitting in a cell that lost its date format — and writes it back consistently into a new column next to the original.
Every fixed value goes into a new column named "<Column> (Fixed)", placed right next to the source. Nothing is deleted, and nothing is guessed at silently — a value that doesn't look like a date at all is left alone, and one that looks like a date but couldn't be parsed is counted and reported rather than skipped without a trace.
Excel stores every date internally as a count of days since 1899-12-30. A cell that's lost its date formatting shows that count as a plain number instead — 45382, for example, is 2024-03-31. This tool spots numbers in a plausible date range and converts them the same as any other date in the column.
A single day-first/month-first choice applies to every ambiguous numeric date in the run — there's no per-row guessing. A date that spells out the month (like "3 April 2024") isn't affected either way.
It's left exactly as it is — the new column is blank for that row, and it isn't counted as an error. Only a value that looks like it's trying to be a date but couldn't be parsed is flagged, with its row number, in the summary.
.xlsx, .xlsm and .csv, up to 10 MB. Three files a month without an account.
It's processed for your session and deleted within the hour. It isn't kept, shared, or used for anything else.