How to Convert Excel to CSV

Converting Excel to CSV should be a two-click job, and most of the time it is. What catches people out is everything that happens quietly on the way: a workbook with five sheets becomes one CSV and nothing warns you the other four are gone, an order number like 00123 arrives as 123, and a column of dates comes out in a format the receiving system rejects.

There are four practical ways to do it — in your browser, in Excel itself, in Google Sheets, and from the command line. This guide covers all four, but the more useful part is the middle: which one to pick based on what is actually inside your spreadsheet, and what each one does to your data on the way out.

Skip to the method that matches your file. The differences are not cosmetic — each one changes your data in a different way.

This is the fastest route when you do not have Excel open, when the workbook has several sheets, or when the data is sensitive. The Excel to CSV converter runs entirely in your browser — the file is read locally and never uploaded to a server.

1. Open the Excel to CSV tool.
2. Drag your .xlsx file onto the drop zone, or click to browse for it.
3. Click Convert.
4. Download the CSV. If your workbook has more than one sheet you get one download per sheet, each named after the sheet it came from.

Two things it does well that desktop Excel does not. First, it exports every sheet: a two-sheet workbook produces two CSV files, where Excel and LibreOffice both give you the first sheet only and say nothing about the rest. Second, text stays text — an order number stored as 00123 comes out as 00123, and a 13-digit product code comes out in full rather than as 9.78012E+12.

Formulas are exported as their calculated results, which is what you want in a CSV: a cell containing a formula that evaluates to 2469 exports as 2469, not as the formula text. Commas and quotation marks inside your data are escaped correctly, so a cell reading a, comma "quote" survives intact.

Two limits worth knowing before you start. It reads the modern .xlsx format only — the legacy binary .xls from Excel 2003 and earlier will fail, so open it in Excel or Google Sheets once and re-save it as .xlsx first. And if your sheet has date columns, read What breaks when Excel becomes CSV below before you rely on the output.

Excel's own export is the right choice when the file contains dates or non-English characters, because Excel writes out what each cell displays rather than the raw value underneath.

1. Open the workbook and click the tab of the sheet you want. Excel exports only the active sheet.
2. Go to File → Save As (or File → Export → Change File Type).
3. In the file-type dropdown choose CSV UTF-8 (Comma delimited) (.csv) — not the plain CSV (Comma delimited) option.
4. Click Save, then OK at the warning that only the active sheet will be saved.

That CSV UTF-8 choice matters more than it looks. The plain CSV option writes the file in a legacy regional encoding, which is how Café Müller ends up in a database as Café Müller. Pick the UTF-8 variant and that class of bug disappears.

The cost of this method is the sheet limit. If you need five sheets out of a workbook you repeat the whole process five times, renaming each file as you go — which is the specific chore the browser tool removes.

Useful when you have no spreadsheet software installed at all, or when the file arrived as an email attachment you would rather not download twice.

1. Go to Google Drive and upload the .xlsx file, then open it in Google Sheets.
2. Select the tab you want to export.
3. Choose File → Download → Comma-separated values (.csv).

Google Sheets writes UTF-8 without asking, which makes it a reliable option for non-English text. It shares Excel's limitation of exporting the active sheet only, and it has a privacy cost the other methods do not: the spreadsheet is uploaded to Google's servers. For anything containing customer records, payroll or medical data, use a method that keeps the file on your own machine.

If you do the same conversion every week, script it. LibreOffice ships a headless converter that needs no GUI and works identically on macOS, Windows and Linux.

soffice --headless --convert-to csv --outdir ./out book.xlsx

To control the delimiter, quoting and encoding explicitly, pass the filter tokens — comma (44), double-quote (34), UTF-8 (76):

soffice --headless \
  --convert-to csv:"Text - txt - csv (StarCalc)":44,34,76 \
  --outdir ./out book.xlsx

On macOS the binary lives inside the app bundle, at /Applications/LibreOffice.app/Contents/MacOS/soffice. Like Excel, it converts the first sheet only — a two-sheet workbook produces exactly one CSV, with no warning that anything was left behind.

For a Python pipeline, pandas is three lines and gives you a place to clean the data on the way through:

import pandas as pd

df = pd.read_excel("book.xlsx", sheet_name="Sales", dtype=str)
df.to_csv("sales.csv", index=False, encoding="utf-8-sig")

Two details there do real work. dtype=str reads every column as text, which is what stops leading zeros disappearing and long IDs turning into scientific notation. encoding="utf-8-sig" writes the byte-order mark that makes Excel on Windows open the file with the right encoding when you double-click it.

A CSV is a plain text file. It has no cell types, no formatting, no formulas, no multiple sheets and no character-set declaration. Everything Excel knows about your data that is not the text itself is discarded at the moment of export — and the losses are silent. These are the five that actually bite.

The rule that explains most of this: CSV export writes what the cell shows, not what it holds. That is worth testing rather than trusting, so I did. The same date, in the same file, converted with the same command, twice — changing only the cell's display format:

So the fix for date columns is not to hunt for an export setting. It is to select the column, set its number format to yyyy-mm-dd in the spreadsheet, and export after that. The same trick fixes leading zeros: format the column as Text before you export, and 00123 stays 00123.

One caveat specific to browser-based converters, including this one. They read the raw stored value rather than the displayed text, which is exactly why leading zeros survive — but it also means a date cell can come out as a full machine timestamp such as Tue Sep 08 2026 17:00:00 GMT-0700 rather than a date, and because that rendering uses your local time zone, a date-only cell can appear to shift by a day. If your sheet has date columns and you need them clean, use Excel or Google Sheets for that file, or convert the date column to text in the spreadsheet first.

Thirty seconds of checking beats a failed import. Open the CSV in a plain text editor — not in Excel, which re-interprets the file as it opens and hides the very problems you are looking for — and check four things.

• The header row is present and the column names are what the receiving system expects.
• Every row has the same number of commas. A row with fewer fields than the header will break a strict parser, and trailing empty cells are the usual cause.
• Accented characters look right. If you see Café instead of Café, the encoding is wrong — re-export as CSV UTF-8.
• Dates and IDs look like you expect. This is where the damage is, and it is invisible from inside Excel.

Going the other way afterwards is straightforward too: CSV to Excel rebuilds an XLSX from your CSV, and CSV to JSON converts it for an API. If you are combining several exports, Merge CSV stacks them into a single file.