Not sure.. but the 101% may come from adding all the logical reads values of the dba_hist_seg_stat (not doing a top N) then divide it by the session logical reads of course they are coming from different sources so there may be instrumentation differences between dba_hist_seg_stat=logical reads and dba_hist_sysstat=session logical reads Just like on the CPU time where you can pull it on different sources like dba_hist_sys_time_model or dba_hist_sysstat or on ASH which is detailed here http://goo.gl/5WaAg -- Karl Arao karlarao.wordpress.com karlarao.tiddlyspot.com -- //www.freelists.org/webpage/oracle-l