This does sound similar to what my developer is talking about... the _optim_peek_user_binds is hidden no? I've never used it. My statistics_level is typical. And the version is 10.2.0.1 Joel Patterson Database Administrator joel.patterson@xxxxxxxxxxx x72542 904 727-2542 ________________________________ From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Charles Schultz Sent: Wednesday, September 13, 2006 9:46 AM To: Wolfgang Breitling Cc: Oracle-L Freelists Subject: Re: "No bind buffers allocated" and "_optim_peek_user_binds = true" I filed a bug with Oracle. Apparently, bind buffers are not peeked when you set statistics_level = ALL in 10.2.0.2. At least, not for my simple tests. I was using statistics_level = all because it makes Cardinality Feedback so much simpler (query 2 dictionary tables, as opposed to ripping apart 10046 and 10053 traces). So much for that idea.... On 9/12/06, Charles Schultz <sacrophyte@xxxxxxxxx> wrote: This is straight pl/sql issued via sqlplus from the command line on the same host. On 9/12/06, Wolfgang Breitling < breitliw@xxxxxxxxxxxxx <mailto:breitliw@xxxxxxxxxxxxx> > wrote: What is the application which shows that? It is possible to for an application to issue a parse call without supplying bind values. There was a problem with a version of JDBC which it didn't send bind info, but I believe that is fixed in 10.2 - assuming the client also has been upgraded to 10.2. Christian Antognini will know more details about this. Quoting Charles Schultz <sacrophyte@xxxxxxxxx >: > Under which circumstances will a bind buffer not be allocated? I was > diagnosing a 10053 trace and even though it looks like the bind variables > should be peeked, I get empty buffers. The query plan seems to reflect this, -- regards Wolfgang Breitling Oracle 7,8,8i,9i OCP DBA Centrex Consulting Corporation www.centrexcc.com -- Charles Schultz -- Charles Schultz