Feather / GeoArrow
Apr 28,2026

GISBox is a one-stop 3D GIS data editing, conversion and publishing platform that supports editing in multiple GIS formats such as OSGB/GEOTIFF/RVT, converting to 3DTiles/Terrain and publishing.

Introduction

Feather / GeoArrow is a geospatial data format extended from Apache Arrow, a columnar data format designed for high-speed data processing. It is designed to efficiently store and transfer large volumes of vector geographic data, achieving faster read/write performance and memory efficiency than conventional GIS formats. It is gaining attention as a next-generation format for high-speed processing of geographic data, particularly in cloud environments and big data analytics platforms. GeoArrow utilizes Arrow's columnar memory structure and is characterized by its ability to manage geometry and attribute data in a unified manner.

File Structure

The data structure of Feather / GeoArrow is built on Arrow's columnar architecture and mainly consists of the following elements:

Schema: Defines the structure of the dataset, describing the type of each field (integer, string, floating point, etc.) and the definition of geometry columns. This allows data structures to be handled uniformly across different environments.
Geometry Column: A dedicated column that stores spatial geometries such as points, lines, and polygons. In GeoArrow, geometry information is efficiently stored using coordinate arrays and offset arrays.
Attribute Columns: Stores attribute data associated with geographic objects (names, classifications, numerical data, etc.) in columnar format. Because of the columnar format, analytical processing and filtering can be performed at high speed.
Metadata: Records spatial information such as coordinate reference system (CRS), geometry type, and encoding method. This improves compatibility between GIS software and data analysis tools.
Feather File Container: A container format that saves Arrow data as files. It stores data in binary format, enabling high-speed reading and writing.

Pros

High-speed data processing performance: Arrow's columnar memory structure enables high-speed execution of data loading, analysis, and transfer. Particularly suitable for processing large geographic datasets.
High memory efficiency: Because data can be efficiently managed in memory, analysis processing performance improves. Additionally, zero-copy processing reduces unnecessary data copying.
High affinity with big data environments: Can work with many data analysis environments such as Python, R, Java, and C++, making it suitable for using geographic data in machine learning and data science fields.
Suitable for cloud and distributed processing: Columnar formats are compatible with cloud data lakes and distributed processing platforms, and can be used as an analysis infrastructure for large-scale GIS data.

Cons

Still limited support in GIS software: Compared to traditional GIS formats (such as Shapefile or GeoPackage), there are still fewer GIS software applications that support it.
Not directly suitable for visualization purposes: GeoArrow is primarily a format for data analysis and processing, so additional conversion may be required when used as a format for map display.
Format specification is still under development: As it is a relatively new technology, specifications may change in the future. Additionally, implementation differences between tools may occur.

Application Scenario

Feather / GeoArrow is suitable for large-scale geospatial data analysis and cloud-based GIS systems. For example, it can be used as an analysis infrastructure for high-speed processing of large volumes of data such as urban traffic data, satellite observation data, and location log data. Additionally, it is increasingly being used as a format for efficiently handling spatial data in data science and machine learning workflows. Furthermore, when combined with cloud data lakes and distributed computing environments, it plays an important role in building next-generation geospatial big data processing infrastructures.