[HUDI-6242] Format changes for Hudi 1.X release line

• Type: Epic
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Fix Version/s: 1.0.1
• Component/s: core
• Labels: hudi-umbrellas
• Epic Name: 1.X Format Changes
• Assignee: Vinoth Chandar
• Reporter: Vinoth Chandar
• Due: 13/Sep/23
• Created: 19/May/23 18:11
• Updated: 5 days ago 00:52

Description

This EPIC tracks changes to the Hudi storage format.

Proposals

Format change is anything that changes any bits related to

  * Timeline : active or archived timeline contents, file names.

• Base Files: file format versions, any changes to any data types, file footers, file names.
• Log Files: Block structure, content, names.
• Metadata Table: (should we call this index table instead?) partition names, number of file groups, key/value schema and metadata to MDT row mappings.
• Table properties: What's written to hoodie.properties.
• Marker files : Can be left to the writer implementation.

Change summary:

The following functionality should be supportable by the new format tech specs (at a minimum)
Flexibility :

• [Pending] Ability to mix different types of base files within a single table or even a single file group (e.g images, json, vectors ...)
• [Pending] Easy integration of metadata for JVM and non-jvm clients (parquet as MT format, HFile native APIs)

Metafields :

• [Resolved] Should _recordkey be uuid special handling?
• Semantics of _hoodie_commit_time , with completion time changes.

Additional Info:

• Support encoding of watermarks/event time fields as first class citizen, for handling late arriving data.
• [Resolved] Position based skipping of base file
• [Pending] Additional metadata to avoid more RPCs to scan base file/log blocks.
• [Pending] ML/Column family use-case?
• [Resolved] Support having changeset of columns in each write, other headers

Log :

• [No change needed] Support writing updates as deletes and inserts, instead of logging as update to base file.
• [Pending] CDC format is GA.

Table organization:

• [Pending] Support different logical partitions on the same data
• [Pending] RFC-60/Storage of table spread across buckets/root folders
• [Pending] Decouple table location from timeline, metadata. They can all be in different places

Concurrency/Timeline:

• [Pending] Ability to support general purpose multi-table transactions, esp between data and metadata tables.
• [Pending] Support lockless/non-blocking transactions, where writers don't block each other even in face of conflicts.
• [Resolved] Support for long lived instants in timeline, break down distinction between active/archived
• [Pending] Support checking of uniqueness constraints, even in face of two concurrent insert transactions.
• [Pending] Support precise time-travel queries
• [Pending] Support time-travel writes.
• [Pending] Support schema history tracking and aid in schema evol impl.
• [Resolved] TrueTime store/support for instant times
• [Pending] No more separate rollback action. make it a new state.

Metadata table :

• Encode filegroup ID and commit time along with file metadata

Table Properties:

• Partitioning information/indexing info

Marker Files:

• Write marker files for logs as well, based on new marker format.