Installer xan avec Homebrew, MacPorts, Nix, pacman, scoop, winget

Liens du projet

• Page d'accueil
• Docs amont
• Page du gestionnaire de paquets

Commandes et alias

• xan

Historique du projet

The project repository was created in 2018, but the public crate history for xan begins with version 0.1.0 on February 7, 2024. The README says xan began as a fork of BurntSushi's xsv and has since been nearly entirely rewritten for Sciences Po medialab use cases rooted in web data collection and social-science analysis.

The tool's development identity is tied to practical research data work: CSV remains a common interchange format among spreadsheets, statistics tools, scripting languages, web archives, and data pipelines, and xan extends beyond plain CSV into formats such as CDX, VCF, GTF, SAM, BED, JSON, Excel, and NumPy arrays.

Historique d'adoption

xan is newer than classic CSV tools, but it has become visible in terminal-data circles because it combines xsv-like speed with a broader command set and interactive exploration features. By July 2026 the GitHub repository had roughly four thousand stars, and crates.io listed tens of thousands of downloads.

Its adoption niche is strongest among users who want shell-native data munging without starting Python, R, pandas, a notebook, or a database engine. That includes journalists, researchers, civic-tech users, data engineers doing quick inspections, and developers manipulating large delimited files in pipelines.

Modes d'utilisation

Typical xan sessions chain commands to inspect headers, preview rows, filter with expressions, compute frequencies, aggregate columns, join files, sort large CSVs, and render quick terminal visualizations. The README highlights its own expression language for cases too complex for the simplest commands but faster and more focused than embedding a general dynamic language.

The package is useful as a one-binary data knife: users can keep transformations in shell history, run them repeatedly in scripts, and work with gigabyte-scale CSVs using streaming, SIMD parsing, and parallelism where commands allow it.

Pourquoi les passionnés de paquets s'y intéressent

xan sits in the lineage of modern Rust CLI rewrites of Unix data tools: fast, single-purpose at the binary level, but broad enough to replace many tiny CSV scripts. Its xsv ancestry matters, but the medialab rewrite and CSV-adjacent commands give it a distinct research-data personality.

Chronologie

• 2018: medialab/xan repository is created.
• 2024-02-07: xan 0.1.0 is published on crates.io.
• 2025: Release cadence accelerates with frequent 0.x releases and expanded command coverage.
• 2026-06-17: crates.io lists xan 0.59.0 as the latest version at the time researched.

Related projects

• xsv is the Rust CSV toolkit from which xan originally forked.
• csvkit, Miller, DuckDB, pandas, and command-line SQL tools overlap with parts of xan's CSV manipulation niche.
• simd-csv is part of the parser story highlighted by xan's README and project blog.

Sources

• https://crates.io/api/v1/crates/xan
• https://github.com/medialab/xan
• https://github.com/medialab/xan/blob/master/docs/blog/csv_base_jumping.md
• https://github.com/medialab/xan/blob/master/docs/cookbook/dataviz.md
• https://github.com/medialab/xan/releases