[IMPALA-13041] Support Reading and Writing Puffin File Stats for Iceberg Tables

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: Backend, Frontend
• Labels: catalog-2024 impala-iceberg
• Epic Color: ghx-label-13
• Assignee: Daniel Becker
• Reporter: Manish Maheshwari
• Created: 26/Apr/24 10:09
• Updated: 22/Jul/24 09:12

Description

Puffin File format is an iceberg upstream spec to support stats for iceberg tables. These stats cannot be both read and used for query planning and written by Impala today. We want to extend support in Impala to do the below - 

1. Read stats from Puffin files 
2. Write stats to puffin files during load/insert/update/delete commands (as applicable)
3. Modify compute stats command for iceberg tables to compute stats and store them in Puffin files