Here's the quick start overview of what will be performed:
- Use default H2 database and setup JNDI reference
- run DDL to create the table to persist stats:
- extend the default MetadataRepository logic, so that the stats will be pulled from the H2 database, above.
- deploy a vdb that fronts the H2 database
- use the Hibernate quick start as an example of providing CRUD access to H2 database
This would demonstrate the how the Metadata Repository can be extended to persist the stats, and provide the Teiid runtime engine access to up-to-date stats.
The demo would allow the person to enter the stats, run a sql test and see the cardinality was being used, by viewing the QueryPlan.
The next step on this would be to add a mechanism for reading the stats from a datasource and updating the stats table. And then providing a scheduler to update the stats on an ongoing basis.
After reading
TEIID-245, not sure how much this quick start will be useful depending on the impact of this jira.