> This works more or less the same in SQLServer so concurrency should not > be an issue as far as the temporary table is concerned, from the docs:
Yes, I believe that concurrency is not an isuue, except that I found (googling) that SQL server is a bit harder to handle - temp tables seem not to survive the procedure while in Oracle the data gets deleted only at COMMIIT. Anyway concurency is not a an issue as rows inserted are visible only in current transaction until commit (and one can just rollback). It's not just a case of advocating someone's workaround, just imagine you'll have to modify/recompile iBatis code with every bug-fix release, oh noooooo ;) GL
