#unicode #spreadsheet #markup #data


USV: Unicode Separated Values (USV) for data markup. This is for spreadsheets, databases, content management, and more. This builds on ASCII Separated Values (ASV) and contrasts with Comma Separated Values (CSV).

31 releases (13 breaking)

new 0.19.1 Apr 10, 2024
0.18.0 Apr 4, 2024
0.15.0 Mar 26, 2024

#165 in Encoding

Download history 16/week @ 2024-02-09 19/week @ 2024-02-16 270/week @ 2024-02-23 172/week @ 2024-03-01 467/week @ 2024-03-08 1615/week @ 2024-03-15 1233/week @ 2024-03-22 896/week @ 2024-03-29 334/week @ 2024-04-05

4,146 downloads per month
Used in 8 crates

MIT OR Apache-2.0 OR GPL-2.0 OR GPL-3.0


Unicode Separated Values (USV) ™

Unicode Separated Values (USV) ™ is a data format that uses Unicode characters for markup.

This USV crate implements the USV specification: https://github.com/sixarm/usv.

This USV crate and aims to help developers build new USV applications, tools, and workflows.

USV characters


  • File Separator (FS) is U+001C or U+241C ␜

  • Group Separator (GS) is U+001D or U+241D ␝

  • Record Separator (RS) is U+001E or U+241E ␞

  • Unit Separator (US) is U+001F or U+241F ␟


  • Escape (ESC) is U+001B or U+241B ␛

  • End of Transmission (EOT) is U+0004 or U+2404 ␄


use usv::*;
let str = "a␟b␟";
let units: Units = str.units().collect();
assert_eq!(units, ["a", "b"]);
assert_eq!(units.into_usv_string(), str);


use usv::*;
let str = "a␟b␟␞c␟d␟␞";
let records: Records = str.records().collect();
assert_eq!(records, [["a", "b"],["c", "d"]]);
assert_eq!(records.into_usv_string(), str);


use usv::*;
let str = "a␟b␟␞c␟d␟␞␝e␟f␟␞g␟h␟␞␝";
let groups: Groups = str.groups().collect();
assert_eq!(groups, [[["a", "b"],["c", "d"]],[["e", "f"],["g", "h"]]]);
assert_eq!(groups.into_usv_string(), str);


use usv::*;
let str = "a␟b␟␞c␟d␟␞␝e␟f␟␞g␟h␟␞␝␜i␟j␟␞k␟l␟␞␝m␟n␟␞o␟p␟␞␝␜";
let files: Files = str.files().collect();
assert_eq!(files, [[[["a", "b"],["c", "d"]],[["e", "f"],["g", "h"]]],[[["i", "j"],["k", "l"]],[["m", "n"],["o", "p"]]]]);
assert_eq!(files.into_usv_string(), str);


The architecture of this crate looks like this, in order of importance:

  • lib.rs: the library entry point.

  • constants.rs: constants for USV characters.

  • token.rs: the USV Token enumerator for returning parser results.

  • iter/: iterators for units, records, groups, files, tokens.

  • style/: style sets of characters for symbols, controls, braces.

  • layout/: layout formats for lines, visual displays, and editors.

  • from/: convert from one thing into another thing.

  • into_usv_string: trait and impl to convert from data into a usv string.

  • examples.rs: data strings suitable for demos and tests.

  • str_ext.rs: string extension traits for parsing USV.

  • svec.rs: a simple macro for creating string vectors.

  • bench/: benchmark tests; this is work in progress.

  • tests/: integration tests placeholder; not needed yet.


A token is the underlying USV enumeration for parsing a string to output:

pub enum Token {

Type aliases

  • Unit = String

  • Units = Vec

  • Record = Units

  • Records = Vec

  • Group = Records

  • Groups = Vec

  • File = Groups

  • Files = Vec

The USV project aims to become a free open source IETF standard and IANA standard, much like the standards for CSV and TDF.

Until the standardization happens, the terms "Unicode Separated Values" and "USV" are both trademarks of this project. This repository is copyright 2022-2024. The trademarks and copyrights are by Joel Parker Henderson, me, an individual, not a company.

When IETF and IANA approve the submissions as a standard, then the trademarks and copyright will go to a free libre open source software advocacy foundation. We welcome advice about how to do this well.


USV is helping us with data projects. We hope USV may help you too.

We welcome constructive feedback about USV, as well as git issues, pull requests, and standardization help.

No runtime deps