[
https://issues.jboss.org/browse/TEIID-2253?page=com.atlassian.jira.plugin...
]
Steven Hawkins commented on TEIID-2253:
---------------------------------------
I tried fleshing out two approaches, detailed below, but neither were very satisfactory.
1. Modify the multi-source metadata to add models representing the sources with light
proxies around the actual metadata. Views with unions were also constructed in the
metadata wrapper. The engine would then try to plan the queries as if they are
partitioned views. This would allow for the possibility of using different
capabilities/metadata for each of the sources, but it still does push multi-source
knowledge into the engine to handle implicitly partitioning joins.
2. Move the effect of multi-source to a processing time concern. While relatively simple
this does not address the poor planning limitations.
The next approach is to treat the source name as a more pervasive pseduo-column available
on all tables/views - even possibly beyond multi-source mode (although we should probably
choose a true reserved term). This would function similar to Oracle's rowid and would
also have the side effect in multi-source mode of driving the join partitioning behavior.
Multi-source issues
-------------------
Key: TEIID-2253
URL:
https://issues.jboss.org/browse/TEIID-2253
Project: Teiid
Issue Type: Enhancement
Components: Query Engine
Reporter: Steven Hawkins
Assignee: Steven Hawkins
The design of the multi-source feature has several issues that need addressed.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see:
http://www.atlassian.com/software/jira