Comments on: SQL Question: Find overlap in periods – using Analytical Functions (LAG and LEAD) Friends of Oracle and Java Fri, 29 May 2015 11:00:21 +0000 hourly 1 By: max Mon, 12 Sep 2005 13:38:39 +0000 /?p=767#comment-2407 when i try your sample (on 9iR2) i get ORA-12054 :o(

By: Toon Koppelaars Sat, 10 Sep 2005 11:58:41 +0000 /?p=767#comment-2406 Lucas, any idea how Oracle serializes this? I’m assuming MV-refresh adheres to standard read-commited
isolation level.
Suppose assignments is emtpy.
t=0 TX1: insert (1, sysdate, sysdate+2);
t=1 TX2: insert (1, sysdate+1, sysdate+3);
t=2 TX1&TX2 both commit;

Is TX2 blocked at t=1?
Or, is commit-processing fully serialized (ie. one starts, fully completes, and only then does the
other one start (which would then fail the check-constraint).