FlatGeobuf

Overview

FlatGeobuf is a performant binary encoding for geographic data based on FlatBuffers. It supports random access via HTTP range requests, making it ideal for streaming large datasets over the web without downloading the entire file.

Key Features

Feature	Description
Binary Format	Compact, fast to parse
Streaming	HTTP range request support
Spatial Index	Optional R-tree for filtered reads
Single File	All data in one .fgb file
Seekable	Random access to features
Attribute Types	Full type support

Advantages

Aspect	FlatGeobuf	GeoJSON	Shapefile
Size	Small	Large	Medium
Parse speed	Fast	Slow	Medium
Streaming	Yes	No	No
Spatial index	Built-in	No	Separate
Single file	Yes	Yes	No

Usage

GDAL/OGR

# convert geojson to flatgeobuf
ogr2ogr -f FlatGeobuf output.fgb input.geojson
 
# with spatial index
ogr2ogr -f FlatGeobuf -lco SPATIAL_INDEX=YES output.fgb input.geojson
 
# query with bbox filter
ogrinfo output.fgb -spat -122.5 37.7 -122.3 37.9

JavaScript (Web)

import { geojson } from 'flatgeobuf';
 
// stream features from URL with bbox filter
const bbox = { minX: -122.5, minY: 37.7, maxX: -122.3, maxY: 37.9 };
const iter = geojson.deserialize('https://example.com/data.fgb', bbox);
 
for await (const feature of iter) {
  console.log(feature);
}

Leaflet Integration

import { geojson } from 'flatgeobuf';
 
async function loadFlatGeobuf(map, url) {
  const response = await fetch(url);
  const iter = geojson.deserialize(response.body);
  
  const layer = L.geoJSON();
  for await (const feature of iter) {
    layer.addData(feature);
  }
  layer.addTo(map);
}

R

library(sf)
 
# read flatgeobuf
features <- sf::st_read("data.fgb")
 
# write flatgeobuf
sf::st_write(features, "output.fgb", driver = "FlatGeobuf")

Python

import geopandas as gpd
 
# read
gdf = gpd.read_file("data.fgb")
 
# write
gdf.to_file("output.fgb", driver="FlatGeobuf")

When to Use

Choose FlatGeobuf when:

• Web streaming of vector data
• Need bbox filtering via HTTP
• Medium datasets (10-500MB)
• Binary efficiency matters
• Building web map applications

Consider alternatives when:

• Analytics queries → GeoParquet
• Human readability → GeoJSON
• Desktop GIS workflows → GeoPackage

Comparison with GeoParquet

Aspect	FlatGeobuf	GeoParquet
Orientation	Row-based	Columnar
Best for	Feature streaming	Analytics
Compression	Moderate	Excellent
Column selection	No	Yes
Ecosystem	Web mapping	Data science

---

Appendix

Created: 2024-12-23 | Modified: 2024-12-23

See Also

• MOC - Geospatial
• Guide - Geospatial File Format Selection
• Row vs Columnar Geospatial Formats
• GeoJSON
• GeoParquet

Backlinks

---

(c) No Clocks, LLC | 2024