How do you query an Iceberg table from ClickHouse?

Use the iceberg() table function with the table's S3 location, or define a named table with the IcebergS3 engine. ClickHouse reads the Iceberg metadata, determines which files it needs, and queries them directly from S3 - no loading step required.

Does ClickHouse copy Iceberg data into its own storage?

No. ClickHouse reads Iceberg tables directly from object storage. The data stays in S3 in open formats like Parquet, and ClickHouse fetches only the files a given query needs.

Why does my ClickHouse Iceberg query fail to reach S3?

Same-region S3 is reachable from an Instaclustr ClickHouse cluster by default. If your bucket lives in a different region than the cluster, you’ll need to enable egress to that region under Integrations -> AWS S3 Region.

What permissions does ClickHouse need to read Iceberg?

Read access only: s3:GetObject and s3:ListBucket on the query role, plus a bucket policy allowing s3:GetObject and s3:ListBucket, scoped to your AWS account with an aws:PrincipalAccount condition.

Why is querying Iceberg from ClickHouse so fast?

Iceberg's manifest files store per-file statistics, so ClickHouse can skip files that can't match a query's WHERE clause before reading any data. Combined with ClickHouse's columnar engine, this keeps most analytical queries in the sub-second range.

Instaclustr Blog

Featured blogs

How to query Apache Iceberg tables from ClickHouse

Learn how to query Apache Iceberg tables from ClickHouse directly in S3 using the iceberg() table function or IcebergS3. No ETL, loading step, or data duplication required.

August 6, 2026

By Walt Ribeiro

Scaling agent systems with Apache Kafka and A2A (Part 4): Visualizing the Agent2Agent protocol

August 4, 2026

By Paul Brebner

Scaling agent systems with Kafka and A2A (Part 3): Agent2Agent Protocol explained

Learn how the Agent2Agent (A2A) protocol works at runtime, including agent discovery, task handling, streaming, push notifications, and Kafka’s role in scalable agent systems.

August 3, 2026

By Paul Brebner

Self-Service Bring Your Own Cloud (BYOC) for AWS: Deploy Managed Clusters in Your AWS Account

Deploy managed clusters in your own AWS account with self-service Bring Your Own Cloud (BYOC) from NetApp Instaclustr — no tickets, no waiting, just faster cloud onboarding.

July 30, 2026

By Mo Ansari

See all featured blogs

How to query Apache Iceberg tables from ClickHouse

August 6, 2026 | By Walt Ribeiro

Learn how to query Apache Iceberg tables from ClickHouse directly in S3 using the iceberg() table function or IcebergS3. No ETL, loading step, or data duplication required.

August 6, 2026 | By Walt Ribeiro