Data
Files that hold information for programs to read: structured text, binary serialisation, databases, scientific and map data, machine learning models, messages and calendars, and the formats that compress files or bundle many files into one.
Groups
Structured text
Data written as readable text, with rules for nesting and quoting.
4 formatsBinary serialisation
Compact binary encodings of structured data, made for programs to exchange.
2 formatsDatabases
Whole databases in a single file: tables, indexes and the records in them.
1 formatScientific
Large arrays of measurements and results, with the metadata that explains them.
1 formatGeospatial
Points, routes and shapes on the Earth, with their coordinates.
2 formatsMachine learning
Trained models: their weights, and in some formats the network that uses them.
3 formatsMessages & calendars
Emails, calendar events and contact cards, each in a plain text format.
3 formatsCompression
One file squeezed into fewer bytes, to be expanded back exactly.
3 formatsArchives
Many files and folders bundled into one, often compressed as well.
4 formatsAll data formats
| Format | Name | About |
|---|---|---|
| Structured text | ||
.CSV | Comma-separated values | A plain text table, one record per line, fields split by commas. |
.JSON | JavaScript Object Notation | Text format for objects, arrays, strings, numbers and booleans. |
.XML | Extensible Markup Language | Nested elements with tags and attributes, written as text. |
.YAML | YAML Ain’t Markup Language | Indented text for configuration: maps, lists and values. |
| Binary serialisation | ||
.MSGPACK | MessagePack | JSON-like values in compact binary, each led by a type byte. |
.PB | Protocol Buffers binary data | Google’s compact binary messages, defined by a .proto schema. |
| Databases | ||
.SQLITE | SQLite database file | A whole SQL database in one file, in pages of B-trees. |
| Scientific | ||
.H5 | Hierarchical Data Format version 5 | Groups and large numeric datasets, organised like a file system. |
| Geospatial | ||
.GEOJSON | GeoJSON | Map points, lines and areas with properties, written as JSON. |
.GPX | GPS Exchange Format | GPS waypoints, routes and recorded tracks, written in XML. |
| Machine learning | ||
.GGUF | GGUF model file | A language model for llama.cpp: metadata and quantised weights. |
.ONNX | Open Neural Network Exchange | A neural network’s graph and weights in one Protocol Buffers message. |
.SAFETENSORS | Safetensors | Model weights after a JSON header, safe to load with no code. |
| Messages & calendars | ||
.EML | Email message | One email as sent: header fields, then a body with MIME parts. |
.ICS | iCalendar | Calendar events, to-dos and time zones as text content lines. |
.VCF | vCard contact | Contact cards: names, numbers and addresses as text lines. |
| Compression | ||
.BZ2 | Bzip2 compressed file | One file compressed with the Burrows–Wheeler transform. |
.GZ | Gzip compressed file | One file compressed with DEFLATE, with a small header and CRC. |
.XZ | XZ compressed file | One file compressed with LZMA2, in indexed, checked blocks. |
| Archives | ||
.7Z | 7z archive | 7-Zip’s archive: solid LZMA2 compression, file list at the end. |
.RAR | RAR archive | WinRAR’s archive, with volumes, recovery records and encryption. |
.TAR | Tape archive | Files in 512-byte blocks, each led by a header; no compression. |
.ZIP | ZIP archive | Files compressed one by one, listed again in a directory at the end. |