[
https://issues.apache.org/jira/browse/PHOENIX-5447?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Geoffrey Jacoby resolved PHOENIX-5447.
--------------------------------------
Resolution: Won't Do
As the CDC project has evolved, we moved away from the idea of a single,
low-level "generic" Avro schema that looked like a WALEdit. While we did
something like this for an internal POC, it can't support a huge array of
Phoenix features, so we're closing this as Won't Do.
CDC streams will have each table have a specific Avro schema that can represent
most (hopefully one day all) Phoenix metadata features.
> Create generic Avro schema for Phoenix Mutations
> ------------------------------------------------
>
> Key: PHOENIX-5447
> URL: https://issues.apache.org/jira/browse/PHOENIX-5447
> Project: Phoenix
> Issue Type: Sub-task
> Reporter: Geoffrey Jacoby
> Assignee: Geoffrey Jacoby
> Priority: Major
>
> For change data capture, it's useful to have both a single generic, low-level
> schema suitable for capturing changes to any and all tables and views, as
> well as richer, higher level schemas suitable for individual tables or views.
> The low level schema will be the <table, [tenantId], [timestamp]> tuple plus
> a collection of Cells. Think of it as an Avro-encoded WAL Entry including the
> annotations from PHOENIX-5435.
> It has two purposes:
> 1. Transport / persistence of low-level change data to entity-specific topics
> for potential backup or migration use cases
> 2. Allowing downstream consumers to "rehydrate" the generic messages into the
> appropriate strongly-typed, higher level schema, optionally supplemented with
> SELECTs to grab unedited fields, and persist to separate high-level
> entity-specific topics for further processing. (The downstream consumer will
> be done in a separate JIRA -- this is just for the schema.)
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
