Not sure to be honest. It *might* help to see TKPROF output for this SQL.
Printable View
Not sure to be honest. It *might* help to see TKPROF output for this SQL.
As far as I have observed, I have had experinces with such SQL going for Hash Aj, where as if put the same code in DML ( like Update and Merge ) it goes for filter.. Sounds like a bug or is that am missing some point..
Abhay.
have you tried raising a TAR?
Once you get past the numpty at 1st line support, might find someone intelligent who could help
Try it on 10.1.0.3, we had reports of plan wierdness in 10.1.0.2 (I think it was with merge cartesian joins though)
After you create those tables, analyze them, then does your merge consistently explain the same way?
This is a good one for Tom Kyte (asktom.oracle.com). He's got a big brain.
Will raise a TAR for sure.. but before i do.. i wanted to make sure I didnt miss anything..Quote:
Originally posted by davey23uk
have you tried raising a TAR?
Once you get past the numpty at 1st line support, might find someone intelligent who could help
Abhay.
All tables are analyzed.. with histograms too..Quote:
Originally posted by KenEwald
After you create those tables, analyze them, then does your merge consistently explain the same way?
This is a good one for Tom Kyte (asktom.oracle.com). He's got a big brain.
BTW, i never get a chnage to post in ASKTOM site .. always see the status as
Quote:
Sorry I have a large backlog right now, please ask a question later
Abhay.
If any one gets a chance to post in Tom Site... can you post this one?Quote:
Originally posted by abhaysk
BTW, i never get a chnage to post in ASKTOM site .. always see the status as
"Sorry I have a large backlog right now, please ask a question later"
Abhay.
Rgds
Abhay.