"read by other session" scenario

I've got an interesting "read by other session" scenario.. and I documented the details here
http://karlarao.tiddlyspot.com/#%5B%5Bread%20by%20other%20session%20scenario%5D%5D that includes the following... workload (CPU,IO,DB), top events, file io, top sqls, ASH report, ASH drill down

on the last part of the document, I got the following questions:

*1) The SQL_ID 7kugzf4d2vbbm*

.. hmm.. i got this bank application (oltp) with intermittent failures on transactions.. then i was focusing on the time of the issue at 12pm-1pm and as per the awr report of that period "update" statement with SQLID 7kugzf4d2vbbm has the top elapsed with high AAS... and i also have the ASH report of the same period but it does not appear there even if I look for v$active_session_history or dba_hist_active_sess_history.. what appears in ASH as the top SQL was the 2nd top on that AWR top sql report with SQLID 8z5wnj8yn5m68...

What's weird for me is that 7kugzf4d2vbbm does not appear in v$active_session_history or dba_hist_active_sess_history at all

I know that the top SQL from AWR gets data from DBA_HIST*SQLSTATS.. and if it's consuming some amount of work, and it's always on the top then it's got to be somewhere in the ASH but it's not..

*2) Read by other session*

How would it be possible that the SQLs on the ASH report waits mostly on this table RCN_DL_RECN_TRXN? And this table is being accessed by *SQL_ID  7kugzf4d2vbbm*... And the table of the SQLs on the ASH report is coming from ISWITCH_TRANSACTIONS...

How does the update operation on RCN_DL_RECN_TRXN affect the table ISWITCH_TRANSACTIONS? or vice versa..

I would appreciate your comments on this...

-- 
Karl Arao
karlarao.wordpress.com
karlarao.tiddlyspot.com

--
http://www.freelists.org/webpage/oracle-l