Excel Changing Numbers to Dates? Stop It and Undo It

Excel is changing numbers to dates because any value shaped like a date, such as 1-2 or 3/14, gets read as one the moment it enters a cell. There is no switch to turn this off. Set the column to Text before the data arrives, and rebuild values already changed with a TEXT formula.

Why is Excel changing numbers like 1-2 into dates?

Excel changes 1-2 into a date because it tries to recognise every entry as a number, a date or text, and "1-2" matches a day-month or month-day pattern. In a US setup it becomes January 2, displayed as 2-Jan. The cell now holds a date serial number, not the characters you typed.

The serial number is the key to understanding the damage. Excel stores every date as a count of days from a fixed starting point in 1900, and the date you see is a format laid over that count. Once 1-2 is stored as the serial for January 2 of the current year, the original hyphen and the order of the digits are gone from the cell. Only the format remembers that it looked like a date.

Microsoft's support page on stopping numbers changing to dates is blunt about the setting: there is no way to turn the behaviour off. What you can do is tell Excel in advance that a column holds text, or mark individual entries so Excel leaves them alone.

Regional settings change which patterns match. A system set to day-month order reads 1-2 as 1 February rather than January 2. That is why the same CSV can open cleanly on one computer and come out full of dates on another.

Statement values that Excel mistakes for dates

On a converted bank statement, the values at risk are short identifiers that use a hyphen or slash. Check numbers, reference codes, store or branch numbers, fraction-style quantities and some account fragments all fit date patterns. Amounts are safe because they have a decimal point, and the real date column is already meant to be a date.

The table below shows example values and what a US-configured Excel does to them. The values are made up for illustration.

Look at the last row. The original had a leading zero, and the rebuild formula has to ask for "mm" to get it back. Choose the format code that matches what the statement printed, not what looks tidy.

Branch or sort codes such as 20-45-67 usually stay as text, because no date has a 45th month or day. That protection is accidental, and a code like 10-11-12 would convert. Treat every hyphenated code column as at risk.

Stopping the date change during a CSV import

The reliable way to stop the change is to import the file with the risky column set to Text. Double-clicking a CSV lets Excel guess every column type, and its guess turns dates-in-disguise into dates. Importing through Data, From Text/CSV lets you set the type before any value lands in a cell.

The steps for Power Query, the importer in current Excel versions:

1. Open a blank workbook. Do not open the CSV directly.
2. Choose Data, then From Text/CSV, and select the file.
3. In the preview, choose Transform Data.
4. Look at the Changed Type step in the Applied Steps list. Power Query adds it on its own and may already have turned codes into dates.
5. Delete that step, or select the reference column and set its type to Text.
6. Set the real date column to Date and the amount column to Decimal Number.
7. Choose Close & Load.

The older Text Import Wizard does the same job. On its third step you click each column in the preview and set the Column data format to Text. Some versions hide the wizard behind a legacy option, and Power Query is easier to repeat.

If a converter offers the output as XLSX instead of CSV, that file arrives with types already set, and Excel does not re-guess them on opening. The risk is mostly with CSV and text files.

Typing reference values so Excel leaves them alone

For a value typed by hand, put an apostrophe in front: '1-2. Excel stores the entry as text and hides the apostrophe. For a block of cells you are about to fill, format them as Text first. Both methods come from Microsoft's own list of workarounds.

The options in more detail:

• Apostrophe. Type '1-2. The cell shows 1-2, and the apostrophe stays hidden. Best for a few entries.
• Text format first. Select the empty cells, press Ctrl+1, choose Text, then type or paste. Formatting cells that already hold dates does not change them back.
• Leading space. Type a space before the value. It works, but the space stays in the cell, and Microsoft warns it can break lookups such as MATCH or VLOOKUP. Use an apostrophe instead.
• Fractions. To enter 1/2 as a fraction, type 0 1/2, with a zero and a space. Excel keeps the value as one half and sets a fraction format.

Pasting is the trap. When you paste text from another program into cells formatted as General, Excel re-reads each value, and date-like codes convert. Set the destination to Text before you paste.

Can you undo a number Excel already turned into a date?

You can rebuild the original in most cases, because the date still encodes the digits you typed. A TEXT formula with the right format code turns January 2 back into "1-2". The rebuild fails when Excel added information you cannot tell apart from the original, such as a year it assumed.

Here is the repair on a column of damaged references:

1. Insert a helper column next to the damaged one.
2. Enter the TEXT formula from the table above that matches the printed pattern, such as =TEXT(A2,"m-d").
3. Fill it down and compare a sample of rows with the PDF.
4. Copy the helper column and paste it over the original as values.
5. Delete the helper column.

The formula returns text, so the result stays safe from further conversion.

Some cases cannot be rebuilt with certainty. If one cell held "1-2" and another "01-02", both may now be January 2, and the formula has no way to know which one had leading zeros. A value like "12-2025" gained a day of 1 that was never printed, so a rebuild must leave the day out. When the pattern varies, go back to the source and reimport the column as Text.

Undo is the quickest fix if you catch it at once. Press Ctrl+Z right after the paste or entry, set the column to Text, and enter the values again.

Real dates and mistaken dates in the same converted column

A column can hold both real dates and mistaken ones when the converter put a reference into the date column, or when wrapped text pushed a code into the wrong row. Fixing the whole column with one formula would wreck the good dates. Flag the strays first, then repair only those.

A quick flag compares each date with the statement period. Any date outside it is suspicious:

=OR(A2<DATE(2026,3,1),A2>DATE(2026,3,31))

Replace the example dates with your statement's first and last day. TRUE marks a row to check by hand against the PDF.

Mistaken dates also cluster on rows that have no amount, which are often wrapped lines or headings, not transactions. Deleting those rows can be part of the same cleanup, and removing blank rows after a PDF conversion covers how to do it without losing real lines.

Real dates that came out wrong, with day and month swapped or the wrong century, are a different problem from codes turning into dates. Fixing converted dates that are wrong in Excel deals with those.

A checklist for reference columns before the next import

The cheapest fix is the one you never need. Run this checklist before you open next month's statement, and date conversion stops being a monthly chore. It takes a minute once you know which columns your statements carry.

• Name the risky columns. List every column that holds codes with hyphens or slashes, such as check numbers and references.
• Import, never double-click. Use Data, From Text/CSV for every CSV.
• Set types explicitly. Text for codes, Date for dates, Decimal Number for amounts.
• Save the query. In Power Query, the steps rerun when you refresh with a new file in the same place.
• Spot-check three rows. Compare a reference, a date and an amount with the PDF before you use the sheet.
• Keep the source. Hold on to the original file so you can reimport a column instead of rebuilding it.

The same habits apply whatever statement you convert, from a bank PDF to an app export; converting a Cash App statement to Excel walks through one such export. If what you have is a PDF someone printed from their own spreadsheet, what comes back when you convert a PDF back to Excel is the place to start.

FastBankConvert is getting ready to open, with output as Excel, CSV or JSON, and it is built to keep dates, descriptions and amounts in separate columns; Get started is where to join.