New Extension: Iceberg

[nlebovits]

Extension Information

• Name: Iceberg
• Prefix: iceberg
• Repository: https://github.com/portolan-sdi/stac-iceberg-extension
• Schema URL: https://portolan-sdi.github.io/stac-iceberg-extension/v1.0.0/schema.json
• Maturity: Proposal
• Scope: Collection

Description

STAC Extension for describing Apache Iceberg table metadata, enabling clients to discover and connect to Iceberg tables for querying geospatial data via PyIceberg, DuckDB, Spark, or BigQuery.

The extension provides metadata fields for:

• Catalog connectivity: iceberg:catalog_type (rest, sql, glue, hive, dynamodb), iceberg:catalog_uri
• Table identification: iceberg:table_id (fully qualified namespace.table)
• Versioning: iceberg:format_version, iceberg:current_snapshot_id
• Partitioning: iceberg:partition_spec (partition fields and transforms)

Designed to complement the Table Extension, which provides schema-level metadata (table:columns, table:row_count). The Iceberg Extension adds catalog connectivity and versioning on top.

Use Case

Geospatial data lakes increasingly store vector data in Iceberg tables (via GeoParquet). This extension allows STAC catalogs to advertise Iceberg table locations so that clients can query data directly rather than downloading bulk files.

Reference Implementation

Portolake generates STAC Collections with Iceberg extension fields from live Iceberg tables via PyIceberg.

Request

Please add ['portolan-sdi', 'stac-iceberg-extension'] to COMMUNITY_REPOS in python/config.py.

Organization: Portolan SDI