Home › Guides › Python

How to convert XLSX to CSV in Python

Convert Excel files to CSV with pandas or openpyxl: one sheet, every sheet, .xls/.xlsb/.ods engines, huge files, and why values differ from what Excel shows.

Updated October 2026

Python is the usual tool for converting Excel files to CSV in scripts and data pipelines. The two common routes are pandas, which is short and handles every sheet at once, and openpyxl, which streams huge files row by row. Both read the values stored in the file, not what Excel displays, which matters for dates, money and leading zeros. Here's how to get it right.

Convert XLSX to CSV with pandas

Install the libraries once:

pip install pandas openpyxl

Then one sheet is two lines:

import pandas as pd

df = pd.read_excel("workbook.xlsx", sheet_name="Orders", dtype=str, keep_default_na=False)
df.to_csv("orders.csv", index=False, encoding="utf-8")

• dtype=str reads every cell as text, so IDs aren't turned into floats like 123.0.
• keep_default_na=False keeps empty cells empty, instead of writing NaN or treating text such as NA as missing.
• index=False stops pandas adding its own row-number column.

Every sheet at once

Pass sheet_name=None to get a dictionary of all sheets:

import pandas as pd

sheets = pd.read_excel("workbook.xlsx", sheet_name=None, header=None, dtype=str, keep_default_na=False)
for name, df in sheets.items():
    df.to_csv(f"{name}.csv", index=False, header=False, encoding="utf-8-sig")

header=None plus header=False copies the rows exactly, without pandas treating the first row specially. utf-8-sig adds a BOM so Excel opens the files with accents intact. If a sheet name contains / or \, clean it before using it as a file name.

.xls, .xlsb and .ods files

pandas picks a reader (an “engine”) by file type, and each needs its own package:

Choose an engine with pd.read_excel("file.xlsb", engine="calamine"). The calamine engine reads all the formats above and is much faster on big files.

Huge files: stream with openpyxl

pandas loads the whole sheet into memory. For files with millions of cells, stream them instead:

import csv
from openpyxl import load_workbook

wb = load_workbook("workbook.xlsx", read_only=True, data_only=True)
for ws in wb.worksheets:
    with open(f"{ws.title}.csv", "w", newline="", encoding="utf-8") as f:
        writer = csv.writer(f)
        for row in ws.iter_rows(values_only=True):
            writer.writerow(["" if v is None else v for v in row])
wb.close()

read_only=True streams rows instead of loading everything, and data_only=True gives the last calculated value of each formula rather than the formula text. newline="" stops Windows adding blank lines between rows.

The catch: values, not what you see

Both pandas and openpyxl return the value stored in the cell and ignore its number format. So:

• A ZIP code stored as the number 501 with a 00000 format comes out as 501, not 00501. (Codes typed in as text keep their zeros.)
• Dates come out as 2026-03-04 00:00:00, whatever style the sheet shows.
• Money and percentages come out as plain numbers: 1249.5 and 0.1.

That's often what you want for a database. If you need the CSV to match what Excel shows, convert with xlsxtocsv.com, which applies each cell's number format, or with LibreOffice from the command line.