[
https://issues.apache.org/jira/browse/IGNITE-24254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17914977#comment-17914977
]
Pavel Tupitsyn commented on IGNITE-24254:
-----------------------------------------
Discussed with [~amashenkov], decided to revert:
* For 3.0, we don't support schemas
* For 3.1 we'll add schemas support, and we want to maintain two-way client
compat (3.0 client can work with 3.1 server and vice versa)
Therefore, we'll keep current TABLE_GET and TABLES_GET operations as is (name
only, no schema), and introduce new ones that work with parsed names.
> Sql schema. Fix Thin client protocol to use schema names
> --------------------------------------------------------
>
> Key: IGNITE-24254
> URL: https://issues.apache.org/jira/browse/IGNITE-24254
> Project: Ignite
> Issue Type: Improvement
> Components: thin client
> Reporter: Andrey Mashenkov
> Assignee: Pavel Tupitsyn
> Priority: Major
> Labels: ignite-3
> Fix For: 3.0
>
> Time Spent: 40m
> Remaining Estimate: 0h
>
> Table can be unambiguously identified either by id or by pair of schema name
> and table name.
> So, Thin Client protocol should be aware of schema name a table belongs to.
> ClientTables, *ClientTableGetRequest* and *ClientTablesGetRequest* should
> either read/write table canonical name (see `QualifiedName.toCanonicalName`)
> or transfer schema name with the table name.
> I'd prefer second approach with 2 separate Strings, to avoid excessive
> parsing and quoting names.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
