[teiid-issues] [JBoss JIRA] (TEIID-1560) Add support for user defined / pushdown aggregate functions

[ https://issues.jboss.org/browse/TEIID-1560?page=com.atlassian.jira.plugin... ] Steven Hawkins commented on TEIID-1560: --------------------------------------- Added end-to-end support with metadata support through index procedures marked as functions the new extension metadata properties (in the namespace uri {http://www.teiid.org/ext/relational/2012}) are: aggregate - to indicate that the function is an aggregate analytic - if the aggregate must be windowed allows-orderby - if the aggregate can be ordered, such as xmlagg allows-distinct - if the aggregate can have DISTINCT specified uses-distinct-rows - if the aggregate effectively uses distinct rows even if distinct is not specified decomposable - if a single argument aggregate function can be decomposed as agg(agg(x)) over subsets of the data To implement a non-pushdown aggregate a developer will implement org.teiid.UserDefinedAggregate with some custom method to act as the accumulator. The accumulator method is effectively treated the same as a normal user defined function, but should return null and will not be static. The parser does not generally support parsing aggregate functions that look like regular functions. For example, "SELECT myagg(x) FROM ..." will not be properly treated as an aggregate and will fail in validation. The workaround is to use ALL - "SELECT myagg(ALL x) FROM ..." or any aggregate specific construct. Confluence updates are still in works and we still need to decide on the DDL support syntax. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.jboss.org/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira