[ https://issues.jboss.org/browse/TEIID-2584?page=com.atlassian.jira.plugin... ] Ramesh Reddy commented on TEIID-2584: ------------------------------------- I have read through both the approaches, seems like dirty flag is improvement on the other technique. I have to say, I did not quite understand the need to invalidating the cache entries first then only refreshing them at time of selection. Other than performance reasons I think the first one of updating single row is sufficient atleast for our needs. Where it follows the pattern {code} create procedure refreshRow(id) AS BEGIN delete from matview where id = $id; select * into matview from mv_original where id = $id; END {code} and whole thing was based on creating triggers on all the tables involved in the view creation. In Teiid, this becomes multi-dimensional problem, where we need triggers on other views/tables in Teiid and on physical source tables. Since there is no good answer, to start with I am fine with extending semantics of nocache hint. The "partial load" script I mentioned is similar, where I was saying user provides the query that updates the mv table. At the same time we can also provide the above single row refresh method. -- 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