Not sure why these tools always want to create their own table. The things become dependent on the version of the Tool rather than the database and then those plans missing the predicate section.

Another bit in support of SQL* Plus being the best tool for all this :)

I guess the 11g documentation is thinking of the case where something like TOAD kicks in and expects to find a real table. I think the best thing to do with utlxplan.sql is to change the code to create a GTT (and make sure it matches the one in catplan.sql!)

Almost all the text say the same thing. But I found this bit in the 11gR2 documentation bit different:

http://download.oracle.com/docs/cd/E11882_01/server.112/e17110/statviews_5146.htm#REFRN29510

While a PLAN_TABLE table is automatically set up for each user, you can use the SQL script utlxplan.sql to manually create a local PLAN_TABLE in your schema.

Yes, either they should retire utlxplan.sql or update it to reflect the changes done in the way plan_table is created.

I forgot to correct the note after Neil’s first comment.
The script that creates the GTT is catplan.sql – but that is part of the 10g installation run, and contains lots of other bits too. The utlxplan.sql script should have been retired by Oracle. but is still lurking, waiting to trap people.

Not sure what I am missing but here I am checking 2 systems (10.2.0.5 on AIX & 10.2.0.4 on HP-UX) and on both the systems utlxplan.sql doesn’t create a global temp table. It is just a simple create table statement. PLAN_TABLE in SYS is created as GTT, though.

Thanks,
Amardeep Sidhu

btw, i solved that issue too

Jithin