RE: Bind values missing from SQL trace
- From: "Rich Jesse" <rjoralist@xxxxxxxxxxxxxxxxxxxxx>
- To: "Allen, Brandon" <Brandon.Allen@xxxxxxxxxxx>
- Date: Wed, 15 Aug 2007 15:46:49 -0500 (CDT)
Hey Brandon,
My only evidence is that trying to duplicate the statement using nulls or
blanks for those variables returns no rows, where the app clearly does. And
some statements that I've seen don't pass a simple smell check -- i.e. the
statement can't possibly return information valuable to the application
given those variables are not populated.
Some binds show up, some don't. But now that you mention it, I don't see
any bind values for character types on queries in this schema (but I do for
other schemas).
I didn't use tkprof because I have six trace files for this one business
process and I wanted the max elapsed between all six. Not that I would have
used tkprof anyway, but this is a good reason for this situation. ;)
Thanks, Brandon!
Rich
> Looks to me like the application is just sending in an empty string for
> the bind variable value. Do you have some evidence to the contrary?
>
> Do other bind variable values show up as expected in the trace files?
>
> Just curious - why use grep to locate the longest-running query - why
> not tkprof with sort=fchela?
>
>
> -----Original Message-----
> From: oracle-l-bounce@xxxxxxxxxxxxx
> [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Rich Jesse
>
>
> Using grep, I quickly locate the longest-running statement, but when I
> view the essential binds, I see this:
>
>
> BINDS #18:
> bind 0:
> value=""
>
> So where's the values for the two binds that I need?
--
//www.freelists.org/webpage/oracle-l
Other related posts:
