This package enables high-performance Excel writing by integrating with the streaming API from the golang package excelize. Users can leverage this functionality without the need to write any Go code, as the entire process can be accomplished through Python.
-
Python and Golang Integration: Seamlessly call Golang built shared libraries from Python.
-
No Golang Code Required: Users can solely rely on Python for Excel file generation, eliminating the need for Golang expertise.
You can easily install the package via pip
pip install pyfastexcelIf you prefer to build the package manually, follow these steps:
-
Clone the repository:
git clone https://github.com/Zncl2222/pyfastexcel.git
-
Go to the project root directory:
cd pyfastexcel -
Install the required golang packages:
go mod download
-
Build the Golang shared library using the Makefile:
make
-
Install the required python packages:
pip install -r requirements.txt
or
pipenv install
-
Import the project and start using it!
| Operating System | Version |
|---|---|
| Linux | Ubuntu 18.04 or higher |
| macOS | macOS 13 (x86-64) |
| Windows | Windows 10 or higher |
- Python 3.8 or higher
For earlier versions of Python or other operating systems, compatibility is not guaranteed.
The index assignment is now avaliable in Workbook and the StreamWriter.
Here is the example usage:
from pyfastexcel import Workbook
from pyfastexcel.utils import set_custom_style
from pyfastexcel import CustomStyle
if __name__ == '__main__':
# Workbook
wb = Workbook()
# Set and register CustomStyle
bold_style = CustomStyle(font_size=15, font_bold=True)
set_custom_style('bold_style', bold_style)
ws = wb['Sheet1']
# Write value with default style
ws['A1'] = 'A1 value'
# Write value with custom style
ws['B1'] = ('B1 value', 'bold_style')
# Write value in slice with default style
ws['A2': 'C2'] = [1, 2, 3]
# Write value in slice with custom style
ws['A3': 'C3'] = [(1, 'bold_style'), (2, 'bold_style'), (3, 'bold_style')]
# Write value by row with default style (python index 0 is the index 1 in excel)
ws[3] = [9, 8, 'go']
# Write value by row with custom style
ws[4] = [(9, 'bold_style'), (8, 'bold_style'), ('go', 'bold_style')]
# Send request to golang lib and create excel
wb.read_lib_and_create_excel()
# File path to save
file_path = 'pyexample_workbook.xlsx'
wb.save(file_path)For row-by-row Excel writing, consider using StreamWriter, a
subclass of Workbook. This class is optimized for streaming large datasets.
Learn more in the StreamWriter documentation.
Explore additional examples in the FullExamples.
The documentation is hosted on Read the Docs.
The following result displays the performance comparison between
pyfastexcel and openpyxl for writing 50000 rows with 30
columns (Total 1500000 cells). To see more benchmark results, please
see the benchmark.
Python keeps the existing workbook and (value, style_name) API. With the
bundled ABI-v2 library it sends JSON metadata plus MessagePack row frames over
a length-aware ctypes boundary; Go decodes one row at a time and writes it with
excelize. Older native libraries are
detected automatically and continue to use the complete JSON/base64 protocol.
The generated XLSX is returned as raw bytes, or written through the compatible
direct-file path when save(path) is used.
Row decoding runs concurrently with excelize's row serialization, workbooks
with multiple stream sheets are written with one worker per sheet, and the hot
Python paths (row_append, the wire encoder) use cached style resolution with
tight per-row loops. Two opt-in speedups are available on top of that:
append_row/append_rows write whole rows per call (2-3x faster than a
per-cell loop, styles may vary per column), and
set_zip_compression_level(6) swaps the final DEFLATE stage for
klauspost/compress (about 3x faster
compression for roughly 20% larger, fully standard files). Without opt-ins the
output archives stay byte-for-byte identical to previous releases.