2.2 Geodatabase Architecture (File, Enterprise, Mobile, and Feature Datasets)

Key Takeaways

  • File Geodatabases (.gdb) are the recommended format for single-user desktop workflows, offering strong performance and no storage limits (aside from 1 TB per dataset).
  • Enterprise Geodatabases run on commercial RDBMS (SQL Server, PostgreSQL, Oracle) and support multi-user editing, versioning, and archiving.
  • Mobile Geodatabases (.geodatabase) are built on SQLite, enabling cross-platform interoperability and offline field data collection.
  • Feature Datasets are logical containers used exclusively to group feature classes that share the exact same spatial reference.
  • Advanced behaviors like Network Datasets, Topology, and Utility Networks must be created inside a Feature Dataset.
Last updated: July 2026

The geodatabase is the core data storage mechanism in the ArcGIS ecosystem. It is much more than a simple file folder; it is an intelligent container capable of storing geometries, attributes, and complex behavioral rules. Understanding the different types of geodatabases and their internal architecture is essential for proper data management in ArcGIS Pro.

Types of Geodatabases

Esri provides several geodatabase architectures to suit different project scales and user needs.

File Geodatabase (.gdb)

The File Geodatabase is the standard, recommended format for single-user desktop GIS. It is stored as a folder of files on the local file system.

  • Storage Limits: There is no limit to the overall size of a file geodatabase. Individual datasets (like a single feature class) can be up to 1 TB by default, and this limit can be configured up to 256 TB.
  • Performance: Excellent performance for rendering and querying.
  • Concurrency: Designed primarily for a single editor. Multiple readers are supported, but only one user can edit a specific dataset at a time (file-locking). It does not support versioning.

Enterprise Geodatabase

Enterprise geodatabases are required for large organizations needing concurrent multi-user editing and central data management. They are built on top of commercial Relational Database Management Systems (RDBMS) such as Microsoft SQL Server, PostgreSQL, Oracle, SAP HANA, or IBM Db2.

  • Storage Limits: Dictated only by the underlying RDBMS hardware and licensing.
  • Concurrency: Supports simultaneous multi-user editing.
  • Advanced Capabilities: Required for complex workflows such as Branch Versioning, Traditional Versioning, and Geodatabase Archiving. These capabilities allow multiple users to edit the same data simultaneously without locking each other out and provide a historical record of all edits.

Mobile Geodatabase (.geodatabase)

The mobile geodatabase is the newest geodatabase format, built on the open-source SQLite relational database.

  • Interoperability: Because it uses SQLite, it is highly portable across different operating systems (Windows, iOS, Android, Linux) without requiring proprietary drivers.
  • Use Case: It is the foundation for offline field data collection in ArcGIS Field Maps and ArcGIS Survey123. It can be fully utilized in ArcGIS Pro for desktop workflows.
  • Limits: The maximum size is 2 TB per database.

[!WARNING] Personal Geodatabases (.mdb), which were based on Microsoft Access, are completely deprecated and are not supported in ArcGIS Pro. You must migrate them to File or Mobile geodatabases.

Internal Geodatabase Architecture

Inside a geodatabase, data is organized into specific structures.

Feature Classes

A feature class is a collection of geographic features with the same geometry type (point, line, polygon), the same attributes, and the same spatial reference. Feature classes can exist directly at the root level of the geodatabase (these are called standalone feature classes).

Feature Datasets

A Feature Dataset is a logical container or folder inside the geodatabase used to organize related feature classes.

The Golden Rule of Feature Datasets: All feature classes placed inside a Feature Dataset MUST have the exact same spatial reference (Coordinate System, X/Y Tolerance, and X/Y Resolution). If a feature class has a different spatial reference, it cannot be imported into the feature dataset.

When to use a Feature Dataset? You should only use feature datasets when absolutely necessary. Their primary purpose is not merely for organizing data like folders on a hard drive. Instead, they are required for implementing advanced geodatabase behaviors that rely on topological relationships between multiple feature classes.

You must use a Feature Dataset to create:

  1. Topology: Rules governing the spatial relationships between features (e.g., "Parcels must not overlap", "Roads must connect at endpoints").
  2. Network Datasets: Used for routing and transportation analysis in the Network Analyst extension.
  3. Utility Networks / Trace Networks: Used for modeling complex connectivity in utilities (water, gas, electric).
  4. Parcel Fabrics: Used for managing land records.

Other Geodatabase Elements

Geodatabases also store:

  • Tables: Standalone attribute tables.
  • Relationship Classes: Define permanent links between two datasets (e.g., linking a parcel feature class to a table of owner records).
  • Domains: Rules that restrict the values allowed in a field (Coded Value domains for dropdown lists, Range domains for numeric limits).
  • Subtypes: Used to categorize features within a feature class (e.g., categorizing a "Roads" feature class into Local, Collector, and Arterial subtypes), allowing different default values and domains for each category.

Comparison of Geodatabase Types

FeatureFile GeodatabaseEnterprise GeodatabaseMobile Geodatabase
Underlying TechProprietary Esri filesRDBMS (SQL, Postgres, etc.)SQLite
Extension.gdbServer connection (.sde).geodatabase
Max SizeUnlimited (1 TB per table)RDBMS limit2 TB
EditingSingle editor per datasetMulti-user concurrent editingSingle editor per dataset
VersioningNoYes (Branch & Traditional)No
Loading diagram...
Geodatabase Hierarchy
Test Your Knowledge

Which of the following is the PRIMARY mandatory requirement for placing multiple feature classes into the same Feature Dataset?

A
B
C
D
Test Your Knowledge

An organization requires ten GIS editors to simultaneously edit a roads dataset and maintain a historical archive of all changes. Which geodatabase type is required?

A
B
C
D
Test Your Knowledge

Which geodatabase format is built on SQLite and is optimized for cross-platform interoperability and offline field data collection?

A
B
C
D