[ https://issues.jboss.org/browse/TEIID-2495?page=com.atlassian.jira.plugin... ] Steven Hawkins commented on TEIID-2495: --------------------------------------- There a a couple of paths here: * detect and rewrite the for each row sql (at least for internally generated procs we should do a better job) * Introduce sql server style instead of triggers, which could be used in addition to for each row and would be a straight-forward way to allow at least bulk insert handling. * Add a general update batching layer to the transaction handling logic The last of which is what I have been pursuing, but it gets quite complicated. Since we are also forcing the instead of blocks to be atomic, we should treat the whole update operation as atomic to prevent a transaction from being created for each row. It also implies that we may want to make the enforcement of atomic in the server somewhat lax to match the expectation when using autoCommitTxn=OFF or in embedded environments without a transaction manager. -- 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