apache / apache/iceberg-python

Databricks Iceberg REST catalog doesn't support multiple snapshots per transaction

Open
#3,076 1 comment 1 reaction 0 assignees View on GitHub
Dominant language
Python
Stars
1.1k
Forks
581
Avg merge
1d 17h
Merged PRs (30d)
77

Description

### Apache Iceberg version

0.10.0 (latest release)

### Please describe the bug 🐞

When connecting with the databricks / unity iceberg REST catalog to write data using pyiceberg, commits fail at `.commit_transaction()` if the transaction contains several snapshots:
`pyiceberg.exceptions.CommitFailedException: CommitFailedException: Failed to commit to the table, requirement failed: Adding multiple snapshots in a single update is not supported.`

This is tricky, since some of the higher-level operations use transactions, but also the "primitives" can fail if a delete happens to contain 2 snapshots (`.update_snapshot()` is called in two places in the delete method).
However these failures might not appear on the happy path but only if eg another writer process has modified the table and the pyiceberg writer then fails because that triggers two snapshots in a transaction

Assuming databricks doesn't intend to bring multi-snapshot transactions into their iceberg rest catalog, would it be possible to add full "autocommit" support to pyiceberg, such that each snapshot can get committed indiviudually?
The transactions would loose their point, but at least one can then again use pyiceberg with unity's iceberg rest reliably

Steps to reproduce:
- Connect to Databricks iceberg REST using pyiceberg
- Open a transaction
- perform operations that generate 2 or more snapshots
- Attempt commit

### Willingness to contribute

- [ ] I can contribute a fix for this bug independently
- [ ] I would be willing to contribute a fix for this bug with guidance from the Iceberg community
- [x] I cannot contribute a fix for this bug at this time

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by reproducing the failure against the Databricks or Unity Iceberg REST catalog with a transaction containing multiple snapshots. Trace .commit_transaction(), the two .update_snapshot() calls in the delete method, and the surrounding transaction behavior; done should mean pyiceberg can reliably commit compatible operations when the catalog rejects multiple snapshots per transaction.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
databases
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.