Repository navigation
Write engine-name, engine-version, and iceberg-version into snapshot summaries #3365
Description
Is your feature request related to a problem or challenge?
A snapshot committed by iceberg-rust records the operation and file metrics, and nothing that identifies the writer. On a table that several engines commit to, that snapshot cannot be distinguished from a client that omits writer identity.
The Iceberg spec defines snapshot summary keys for identifying the engine that wrote the snapshot:
| Key | Description |
|---|---|
engine-name |
Name of the engine that created the snapshot |
engine-version |
Version of the engine that created the snapshot |
iceberg-version |
Version of the Iceberg library that created the snapshot. Not a spec field; Java and Go write it |
Other implementations already write these, or part of them:
- Iceberg Java and Iceberg Go copy an environment context into the snapshot summary. That includes
iceberg-version, and an engine can also setengine-nameandengine-version. apache/iceberg#9273, apache/iceberg-go#1680 - PyIceberg writes
engine-nameandengine-version. apache/iceberg-python#3441
Describe the solution you'd like
On every snapshot iceberg-rust commits, set iceberg-version to Apache Iceberg Rust <crate version>. This replaces a caller value.
Do not default engine-name or engine-version. Those name the embedding engine, which this library does not know. A caller can still set them through set_snapshot_properties, and those values are left unchanged. This matches Java and Go, whose default environment context contains only iceberg-version.
Other caller properties are unchanged. Existing snapshots are left alone.
Willingness to contribute
I can contribute to this feature independently
Activity