csvcut

csvcut is a fast, header-aware command-line tool for slicing CSV files, in the spirit of cut. Select or remove columns, filter and deduplicate rows, then export the result as a clean table, Markdown, HTML, JSON, Excel XML, or CSV. It is written in Zig and ships as a single binary with no runtime dependencies.

Features

• Column selection - choose columns by name or 1-based index, use ranges such as --include Name,3-5, or drop columns with --exclude.
• Row filtering - match fields exactly (Department=Engineering) or with regular expressions (Name~^A) via --filter, and exclude matching rows with --filterOut.
• Deduplication and counting - collapse repeated rows with --unique or summarize them with --count, which adds a count column.
• Multiple output formats - CSV, compact or evenly spaced Markdown, Jira/Confluence tables, console tables, plain HTML, an interactive Handsontable HTML page, JSON as objects or arrays, and Excel XML.
• Flexible input - comma, semicolon, pipe, or tab separators; optional single or double quoting; headers from the file, from the command line, or no header at all; fixed-width records via --lengths; and skipped lines.
• Limits and cleanup - trim surrounding whitespace with --trim, skip lines with --skipLines, cap input with --inputLimit, and cap output with --outputLimit.
• Multiple files - process several input files independently in one invocation, writing to stdout or to a file with -o.
• Config files - a default.config in the current directory is applied automatically, and named configs can be loaded with --config; $CSVCUT_CONFIG overrides the config directory.
• Fast by default - memory-mapped input, no quote processing unless requested, and a compact native executable.
• Helpful diagnostics - --listHeader, --time, and a documented list of exit codes via --exitCodes.

Usage

csvcut [OPTION]... FILE...

Each FILE is processed independently. Formats that emit a complete document (JSON, HTML, Excel XML, and their variants) produce one document per input file. Input from stdin is not supported.

Run csvcut --help for the full option reference, which is also kept in src/USAGE.txt. A quick example:

csvcut people.csv --include Name,Department --filter Department=Engineering --format table

Installation

Prebuilt binaries for Linux (x86_64 and aarch64) and Windows (x86_64) are available on the releases page. Download the binary for your platform and add it to your PATH. No installation or runtime dependencies are required.

Build requirements

Building csvcut only requires the Zig compiler, which can be downloaded from https://ziglang.org/download/. Currently Zig 0.17.0 is supported; builds may break with newer or older versions.

Running zig build test additionally requires bash and python3 for the integration tests.

Windows Zig setup example for x86_64

Execute the following commands in a Windows Command Prompt (cmd.exe):

curl https://ziglang.org/download/0.17.0/zig-x86_64-windows-0.17.0.zip --output zig.zip
tar -xf zig.zip
del zig.zip
set PATH=%cd%\zig;%PATH%

Linux Zig setup example for x86_64

Either install Zig 0.17.0 with your package manager or execute the following commands in a shell:

wget https://ziglang.org/download/0.17.0/zig-x86_64-linux-0.17.0.tar.xz
tar -xf zig-x86_64-linux-0.17.0.tar.xz
rm zig-x86_64-linux-0.17.0.tar.xz
export PATH=$(pwd)/zig-x86_64-linux-0.17.0:$PATH

Build

With Zig installed and on your PATH, run:

zig build

The first build takes a while. When it finishes, the executable (csvcut or csvcut.exe) is in zig-out/bin/.

For an optimized release build for just your native platform, use zig build native. This saves a compressed executable as zig-out/bin/csvcut (csvcut.exe on Windows).

To build compressed release binaries for all supported platforms, run:

zig build release

This saves the following files to zig-out/bin:

• csvcut-x86_64-linux
• csvcut-aarch64-linux
• csvcut-x86_64-windows.exe

Tests

Run the built-in unit and integration tests with:

zig build test

License

csvcut is licensed under the MIT license. See LICENSE.txt.