sheets — A high-performance C-backed CSV engine for Lua

[Shivam Sharma]

Hello Lua community,

I'm excited to share Sheets, a high-performance CSV library for Lua built with the goal of bringing modern, production-grade data tooling to the Lua ecosystem.

Lua is one of the most elegant languages I've worked with: minimal, fast, embeddable, and extremely flexible. However, I have often felt that the ecosystem lacks robust tooling in areas like data processing, analytics, and modern developer workflows.

Sheets is my first step toward helping improve that.

Unlike pure-Lua CSV parsers, Sheets uses a hybrid architecture:

- C backend powered by libcsv for fast parsing and writing

- Clean Lua API inspired by Python's csv module

- Cross-platform support via LuaRocks

==================================================

KEY FEATURES

==================================================

- High-performance CSV parsing using libcsv

- Python-style API:

    * reader

    * writer

    * DictReader

    * DictWriter

- CSV dialect support:

    * Excel

    * Excel-Tab

    * Unix

    * Custom dialects

- CSV Sniffer (delimiter detection)

- Utility functions:

    * read_csv

    * write_csv

    * parse

    * filter

    * sort

    * group_by

    * transpose

    * column extraction

- Cross-platform support

==================================================

COMPARISON WITH PYTHON CSV MODULE

==================================================

Python:

    import csv

    with open("data.csv") as f:

        reader = csv.reader(f)

        for row in reader:

            print(row)

Lua (Sheets):

    local sheets = require("sheets")

    local f = io.open("data.csv")

    for row in sheets.reader(f) do

        print(row[1], row[2])

    end

    f:close()

Dict-based reading:

    for rec in sheets.DictReader(io.open("data.csv")) do

        print(rec.name, rec.age)

    end

==================================================

INSTALLATION

==================================================

    luarocks install sheets

==================================================

PROJECT LINKS

==================================================

GitHub Repository:

https://github.com/NNEngine/sheets

LuaRocks Package:

https://luarocks.org/modules/nnengine/sheets

Documentation Website:

https://nnengine.github.io/sheets/

==================================================

TESTED ON

==================================================

- Windows

- Linux (WSL / native)

- Termux (Android)

==================================================

VISION

==================================================

This is only the beginning.

My broader goal is to help build serious tooling for Lua in areas such as:

- data processing

- analytics

- numerical computing

- machine learning infrastructure

I believe Lua has tremendous potential beyond scripting and embedding, and I hope projects like this can help strengthen the ecosystem.

I would greatly appreciate feedback, suggestions, bug reports, or contributions from the community.

Thank you.

Best regards,

Shivam Sharma

GitHub:

https://github.com/NNEngine

[Thijs Schreijer]

On Sat, 4 Jul 2026, at 15:22, Shivam Sharma wrote:

Hello Lua community,

I'm excited to share Sheets, a high-performance CSV library for Lua built with the goal of bringing modern, production-grade data tooling to the Lua ecosystem.

Nice! I don't have a lot of csv parsing to do these days, but will definitely keep this one in mind for my next csv adventure.

Thijs

[Marc Balmer]

Also, csv is a bit boring. Why not go for the open document format?

Am 04.07.2026 um 19:14 schrieb 'Thijs Schreijer' via lua-l <lu...@googlegroups.com>:



--
You received this message because you are subscribed to the Google Groups "lua-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lua-l+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/lua-l/e704aa41-c7f5-47f4-b847-2e6269f00e36%40app.fastmail.com.

[Martin Eden]

On 2026-07-04 19:16, Marc Balmer wrote:
> Also, csv is a bit boring. Why not go for the open document format?

CSV is real-world.

I remember writing CSV parser in Lua like ten years ago.
Input was 700 MB .csv from library Genesis.

I think OP will discover how dirty real data can be (wrong quoting
and different format flavors instead of one standard). And how hairy
processing code becomes of this.

And I won't dare to do it in ancient language without variable-length
structures like C.

-- Martin

[Foster Schucker]

I worked with a 117,000 rows with all the employee data on a migration
from System A -> B.  A could only produce CSV and we spent a week
cleaning data.  I had an AWK hand written parser we used that did all
the changes.  The final AWK change once file was over 3K in AWK code
plus the parser.

Fun times??