Apache Iceberg input

Read an Apache Iceberg table from AWS Glue, S3 Tables or a REST catalog, with time-travel and column pruning. Its columns show without a run.

Category

Input

Connection

Apache Iceberg

Edition

Enterprise

What it does

Reads an Apache Iceberg table from the AWS Glue Data Catalog, Amazon S3 Tables or an Iceberg REST catalog through a saved Connection and streams it into the pipeline. Pick a namespace and table with the live browser. Because an Iceberg table carries its own schema, its columns show in the builder WITHOUT a run. Optional time-travel reads an older snapshot, and a column list prunes the read.

Inputs and outputs

When to use it

• Read a lakehouse table your catalog already tracks: in AWS Glue, S3 Tables, or a REST catalog such as Apache Polaris
• Reproduce an old result by reading a past snapshot (time-travel)
• Feed downstream nodes real columns with no crawler and no run

Examples

namespace: sales · table: orders

id, customer, amount

As of: 2026-07-01T00:00:00Z

Snapshot id: 3051729675574597004

Tips

• Storage credentials for the data files come from the Connection: its own access key, short-lived credentials a REST catalog vends, or, on a shared connection, the server's IAM role. For Glue and S3 Tables the same key also signs the catalog calls. They are never in the pipeline.
• A snapshot id wins over an as-of time; blank both to read the current snapshot.
• No crawler needed: the table's own metadata already knows its schema, partitions and files.

Related nodes