Transparent Archiving for Partition-by-Range (PBR) Tablespaces

Feature Proposal: In-Place Transparent Archiving for PBR Tablespaces 

Current transparent archiving frameworks require duplicating table structures, mapping base-to-archive targets, and running resource-intensive DELETE routines to purge old data. This proposal introduces an In-Place Metadata Archiving model for Partition-by-Range (PBR) tablespaces. Instead of moving data, users can directly flag existing historical partitions as "archived" within the native table structure. This eliminates DDL duplication, slashes transaction logging overhead, and automates data lifecycle management.

The Problem

• High Overhead: Moving data from active PBR tables to separate archive tables consumes massive CPU, storage bandwidth, and transaction log space.
• Complex Management: DBA teams must maintain duplicate schemas (LIKE tables) and manage complex partition-splitting workarounds.

Proposed Solution: In-Place Partition Flagging

Instead of treating archival as a data relocation task, this feature treats archival as a partition state property. Inside a PBR table, specific range partitions can be flagged directly as transparently archived.

Concrete Example: 7-Year Rolling Active Window

Consider an enterprise table tracking data from 2010 to 2026 partitioned by year. To maintain peak performance, the business only requires the last 7 years (2020–2026) to be fully active for daily operations.

Instead of exporting and deleting the 2010–2019 data, the DBA runs a single command to flag the older decade as an archive