You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@phoenix.apache.org by "Maryann Xue (JIRA)" <ji...@apache.org> on 2015/11/14 19:23:11 UTC

[jira] [Commented] (PHOENIX-2416) Implement multi-tenant tables in Phoenix/Calcite integration

    [ https://issues.apache.org/jira/browse/PHOENIX-2416?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15005521#comment-15005521 ] 

Maryann Xue commented on PHOENIX-2416:
--------------------------------------

[~jamestaylor], [~julianhyde], we were thinking to model multi-tenant tables as views and to make tenantId a system function call. But I just realized that there might be a problem here, for tenantId (although a fixed value for each connection) can be of different data types to different multi-tenant tables. For example, table A defines its tenant_id column as VARCHAR while table B defines its tenant_id column as INTEGER, thus the tenantId() function should have different return types for these two tables.
An alternative is to create ad-hoc views for each multi-tenant table in PhoenixSchema for a multi-tenant connection, and Phoenix metadata does not have to know about these virtual views.
Thoughts?

> Implement multi-tenant tables in Phoenix/Calcite integration
> ------------------------------------------------------------
>
>                 Key: PHOENIX-2416
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-2416
>             Project: Phoenix
>          Issue Type: Bug
>            Reporter: Maryann Xue
>            Assignee: Maryann Xue
>




--
This message was sent by Atlassian JIRA
(v6.3.4#6332)