[
https://issues.apache.org/jira/browse/PHOENIX-3533?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16511392#comment-16511392
]
Geoffrey Jacoby commented on PHOENIX-3533:
------------------------------------------
The problem with doing this is that it potentially breaks any schema management
our users are doing. For example, if an operator wants to spin up a second
cluster, and does so using the exact DDL that created the first cluster, it
would either fail, or create a mutable table, which isn't what the operator
wants.
Any reason why the engine can't read IMMUTABLE_ROWS=true from the DDL and
transparently rewrite the query to CREATE IMMUTABLE TABLE?
At the very least, I'd think we'd want a clear deprecation path, similar to
what HBase does where an API is publicly deprecated in one release and then
removed in the following.
Also, if CREATE IMMUTABLE TABLE exists (and isn't just a planned feature), the
docs don't mention it.
> Remove IMMUTABLE_ROWS TableProperty
> -----------------------------------
>
> Key: PHOENIX-3533
> URL: https://issues.apache.org/jira/browse/PHOENIX-3533
> Project: Phoenix
> Issue Type: Task
> Reporter: Thomas D'Silva
> Priority: Major
> Fix For: 4.15.0
>
>
> Using CREATE IMMUTABLE TABLE ... should be the only way to specify a table is
> immutable, and this property cannot be altered.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
