[BEAM-8367] Python BigQuery sink should use unique IDs for mode STREAMING_INSERTS

• Type: Bug
• Status: Resolved
• Priority: P0
• Resolution: Fixed
• Fix Version/s: 2.17.0
• Component/s: sdk-py-core
• Assignee: Pablo Estrada
• Reporter: Chamikara Madhusanka Jayalath
• Created: 08/Oct/19 18:33
• Updated: 16/May/20 14:20
• Resolved: 23/Oct/19 18:27

Description

Unique IDs ensure (best effort) that writes to BigQuery are idempotent, for example, we don't write the same record twice in a VM failure.

 

Currently Python BQ sink insert BQ IDs here but they'll be re-generated in a VM failure resulting in data duplication.

https://github.com/apache/beam/blob/master/sdks/python/apache_beam/io/gcp/bigquery.py#L766

 

Correct fix is to do a Reshuffle to checkpoint unique IDs once they are generated, similar to how Java BQ sink operates.

https://github.com/apache/beam/blob/dcf6ad301069e4d2cfaec5db6b178acb7bb67f49/sdks/java/io/google-cloud-platform/src/main/java/org/apache/beam/sdk/io/gcp/bigquery/StreamingWriteTables.java#L225

 

Pablo, can you do an initial assessment here ?

I think this is a relatively small fix but I might be wrong.

Issue Links

• links to: GitHub Pull Request #9797