When you use Excel's Save As CSV, you keep only the values on the active sheet, written as plain text. Formulas become their results, and formatting, charts, comments and every other sheet are left behind. Leading zeros and long numbers need care before saving. For most uses, pick CSV UTF-8 in the Save as type list.
Only the active sheet survives Excel save as CSV
A CSV file holds one table, so Excel saves only the sheet that is selected when you save. Every other tab in the workbook is left out of the file. Excel warns about this in a dialog, but the warning is easy to click past, and the missing sheets are not mentioned again.
Microsoft's list of file formats supported in Excel says each CSV type "saves only the active sheet". Microsoft's guide to importing and exporting text files adds that a dialog reminds you only the current worksheet will be saved, and that other sheets need separate saves.
The consequence for a workbook with several sheets is a choice:
- Save each sheet separately. Select the sheet, save as CSV with a name that says which sheet it is, and repeat. Each CSV gets one sheet.
- Combine first, then save. If the sheets hold the same kind of rows, such as one sheet per month of transactions, stack them on one sheet and save that sheet once.
Workbooks made from multi-page PDFs often have this shape, with one sheet per page or per table. Stacking those pages into one table before saving gives one clean CSV instead of a folder of fragments. Extracting multiple tables from a PDF covers how to decide which tables belong together before you get to this step.
One more trap sits here. After saving, the workbook on screen is now the CSV, named with a .csv extension. Keep working in it and close it, and the other sheets are gone for good unless you saved the original workbook too. Save the XLSX first, then save the CSV copy.
Formulas, formatting and charts in a workbook saved as CSV
A workbook saved as CSV keeps only what a cell shows as text. Formulas are replaced by their current results, number formats collapse into plain values, and anything that is not a cell value, such as charts, colours, borders, comments and column widths, is dropped. Microsoft's format list puts it briefly: saving in any text format loses all formatting.
The table below shows what each part of a typical workbook becomes. Example values are for illustration.
| In the workbook (example) | In the saved CSV |
|---|---|
A formula =SUM(C2:C40) showing 1,284.50 |
The value only, no formula |
| A date formatted as 14-May-2026 | The date as text, in a form that depends on its format and your settings |
| An amount formatted as $1,250.00 | A number, possibly with symbols, depending on the format |
| Bold headings, colours, borders | Nothing |
| Column widths, frozen panes, filters | Nothing |
| A chart or picture | Nothing |
| Cell comments and notes | Nothing |
| A second sheet | Nothing; see the previous section |
| Merged cells | The value in the first cell, blanks in the rest |
The formula row matters most for anyone who will open the CSV again. The total is now a fixed number. If someone edits an amount in the CSV, the total does not change. If the CSV is meant for another program, delete total rows before saving, because an importing app reads them as one more record.
Dates and currency deserve a look before saving. A date written in a long display format can arrive in the CSV as text that the next program cannot read as a date. Set date columns to an unambiguous format, such as year-month-day, and amount columns to a plain number format with no currency symbol and no thousands separator, then save.
What happens to leading zeros and long numbers in an Excel CSV?
Leading zeros and long numbers are usually lost before the save, not during it. If a cell holds the number 2134 rather than the text 02134, the CSV gets 2134. If a 16-digit reference was stored as a number, Excel already cut it to 15 significant digits, and the CSV gets the cut version. Fix the cells first.
Microsoft's article on keeping leading zeros and large numbers explains both behaviours. Excel removes leading zeros when it treats a value as a number, and it keeps at most 15 significant digits, so card numbers and other 16-digit identifiers need to be stored as text.
What to check before you save:
- Zero-padded identifiers. ZIP codes, check numbers and branch codes should be Text cells showing their zeros.
- Long identifiers. Card numbers, transaction IDs and some account numbers should be Text cells showing every digit.
- Scientific notation on screen. A column showing 1.23E+15 can reach the CSV in that short form. Widen the column and set a number or text format so the full value shows.
- Numbers stored as text. The reverse case: amounts that are Text cells save fine but may keep stray spaces or symbols. Clean them so the next program reads them as numbers.
There is a matching risk at the other end. When anyone opens the CSV by double-clicking, Excel guesses the types again and can strip the zeros you kept. That is not a fault of the file. Tell the recipient to import it with those columns set to Text.
Which CSV type should you pick in Excel's Save as list?
Pick CSV UTF-8 (Comma delimited) unless the receiving program says otherwise. UTF-8 stores every character, so accented names and symbols survive. The other choices, CSV (Comma delimited), CSV (Macintosh) and CSV (MS-DOS), use older single-byte encodings or line endings meant for particular systems, and some characters do not survive them.
What each one means in practice:
| Save as type | Use it when |
|---|---|
| CSV UTF-8 (Comma delimited) | The default choice for modern apps, databases and scripts |
| CSV (Comma delimited) | An older program says it needs the Windows encoding |
| CSV (Macintosh) | A legacy Mac program asks for it, which is now rare |
| CSV (MS-DOS) | An old DOS-era system asks for it, which is also rare |
The delimiter is a separate issue from the type. In regions where the comma is the decimal mark, Excel may write semicolons between fields, because it follows the list separator set in the system's regional settings. Microsoft's text-file guide notes that changing that setting in Windows affects every application on the computer, so it is better to set the delimiter in the receiving program than to change it system-wide.
If the receiving program shows garbled letters such as "Café" after you save as UTF-8, the program is reading the file with the wrong encoding. Look for an encoding setting on its import screen and choose UTF-8; why accents turn to junk in a CSV explains the pattern and the fixes.
Reopening the saved CSV before you send it
Check the saved CSV in a plain text editor before you send it, not by reopening it in Excel. A text editor shows exactly what the file holds. Excel reinterprets the file as it opens it, which hides problems and can introduce new ones, and saving again from that view can damage values that were fine.
A short pre-send checklist:
- Open the CSV in Notepad, TextEdit or another text editor.
- Confirm line one is the header and the number of lines matches the rows you expected.
- Look at a row with a comma in a text field. That field should be wrapped in double quotes.
- Check one identifier with leading zeros and one long number for every digit.
- Check one date and one amount for the format the recipient expects.
- Look for formula results that should not be there, such as total rows.
The re-save trap is worth spelling out. Open a CSV by double-clicking, and Excel turns 02134 into 2134 and a long reference into a rounded number. Save it, even to fix a single typo, and those changes are written into the file. Make edits in the original workbook instead, then save a fresh CSV.
A CSV that will be opened in Excel by someone else has one more risk: a cell that starts with an equals sign or a similar character can be read as a formula when the file opens. That is called CSV injection, and CSV injection and how to defuse it explains which leading characters to watch for in text fields.
When saving a workbook as CSV is the wrong export
Saving as CSV is the wrong export when the person or program on the other end needs more than one flat table. If they need several sheets, working formulas, formatting or exact column types, send the XLSX workbook instead. A CSV is right only when the next step is an import that wants plain rows.
Signs that XLSX is the better choice:
- The recipient will work in Excel. They lose nothing by getting the workbook, and they avoid the type guessing a CSV triggers when opened.
- The data spans sheets that belong together. A summary and its detail rows stay linked in one workbook.
- Formulas need to keep working. A forecast or a reconciliation template is useless as fixed values.
- The layout carries meaning. Highlighted rows, comments and grouped sections all vanish in a CSV.
The reverse journey has its own steps. Bringing a CSV back into Excel without losing types is an import, not an open, and converting CSV to Excel walks through it.
For bank statements, the choice between the two often comes before Excel is involved at all. FastBankConvert, which is getting ready to open, is built to produce either an Excel file or a CSV straight from the statement PDF, so you can pick the format the next step needs. Get started puts you on the list for trial conversions.