We'll thats exactly what I'm doing right now, studying Oracle University
instructor guides to temporarily start thinking like OU myself again -
I'll
be instructing an OCP Review course next week, meant for people who want
to
pass OCP. And in order to not distract the students, I won't even
After the students have scribbled everything down, he then leans forward
and
quietly whispers to them that in the real world, heart surgeons actual
first check whether or not it's actually necessary to cut out the heart
*beforehand*. He then gives them a little wink and a nod, the students
What I have heard is that all the OCP questions are taken from the Oracle
University Student Guide. After all, you wouldn't you expect the class to
prepare you? Someone suggested that you think like a computer. Well, for
the philosophy behind the exam, think like an organization, namely
Oracle
Btw, why do you want to take an errorstack on a behaviour-changing event?
Tanel.
- Original Message -
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Sent: Tuesday, January 06, 2004 11:24 PM
Oracle doesn't but Jonathan Lewis does, in his Tutorials. I found out
about that
Hello:
I was looking through some OCP questions posted on the web and came across
the one below.
I believe the answer is (D), because the join type would be dependent on the
number of rows within the table. Is this correct or does the OPTIMIZER_MODE
set to FIRST_ROWS alter this behavior?
The
D is probably the answer, but most of the time oracle will choose nested
loop.
Beware the OCP tuning test. It is completely and totally inaccurate. I
emailed the author of the Sybex tuning book and he agreed with me. He said
he wrote the book to the test and knows its garbage.
- Original
Nope. The answer is b). In the FIRST_ROWS mode, optimizer prefers NL to all other
methos despite the price.
On 2004.01.06 13:44, Jay Wade wrote:
Hello:
I was looking through some OCP questions posted on the web and came across
the one below.
I believe the answer is (D), because the join
Tuesday, January 6, 2004, 1:59:26 PM, Mladen Gogala ([EMAIL PROTECTED]) wrote:
MG Nope. The answer is b). In the FIRST_ROWS mode, optimizer prefers NL to all other
MG methos despite the price.
Does Oracle themselves *document* that what you say is the
case? I believe you, but I'm not sure that
thought so, I'm not 100% certain the OCP will say that though. alot of
inaccuracies in that test.
btw, Ive been playing with first_rows lately. I've noticed that it has a
preference for 'INDEX FULL SCAN' over 'INDEX RANGE SCAN'. Ive found that in
some test cases where you have two tables
Think like a computer.
Which execution plan will be the result?
result of what?
an insert statement?
--
Please see the official ORACLE-L FAQ: http://www.orafaq.net
--
Author: [EMAIL PROTECTED]
INET: [EMAIL PROTECTED]
Fat City Network Services-- 858-538-5051 http://www.fatcity.com
San
Note in-line.
Regards
Jonathan Lewis
http://www.jlcomp.demon.co.uk
The educated person is not the person
who can answer the questions, but the
person who can question the answers -- T. Schick Jr
One-day tutorials:
http://www.jlcomp.demon.co.uk/tutorial.html
Three-day seminar:
see
FIRST_ROWS would alter the behavior regardless of the number of rows.
-Original Message-
Jay Wade
Sent: Tuesday, January 06, 2004 1:44 PM
To: Multiple recipients of list ORACLE-L
Hello:
I was looking through some OCP questions posted on the web and came across
the one below.
I
What I have heard is that all the OCP questions are taken from the Oracle
University Student Guide. After all, you wouldn't you expect the class to
prepare you? Someone suggested that you think like a computer. Well, for
the philosophy behind the exam, think like an organization, namely Oracle
Jonathan, you're right. Interesting thing is that bitmap indexes, which were made for
DW processing and not for OLTP will also be considered for NL context in First_Rows
mode.
Here is the proof, which also proves that I'm a lousy typist:
SQL set autorace on explain
SP2-0158: unknown SET option
What I meant is that the question cannot be answered without making human
assumptions about the question itself. It is a little difficult (Note
little not lot) to believe that such a poorly written question would
appear on a test that costs money to take.
-Original Message-
Someone
Oracle doesn't but Jonathan Lewis does, in his Tutorials. I found out about that
from Scott Gosset in 8i internals class in NYC. Seems still to be true.
Gospel of Jonathan should suffice, however. I just discovered that my
10053 trace name errorstack forever, level 12 causes ORA-600 in 9.2. Let
recipients of list ORACLE-L [EMAIL PROTECTED]
Subject: Re: OCP Question (Perf Tuning)
Date: Tue, 06 Jan 2004 11:34:26 -0800
thought so, I'm not 100% certain the OCP will say that though. alot of
inaccuracies in that test.
btw, Ive been playing with first_rows lately. I've noticed that it has a
preference
Jonathan noted that
Nope. The answer is b). In the FIRST_ROWS mode, optimizer
prefers NL
to
all other
methos despite the price.
Unless the alternative is a full tablescan on the inner
table - in which case merge or hash joins can be
considered.
For some reason hash joins were
I see you're running on Oracle 9 there, and that
can make a big difference. After posting my
hypothesis, I created a test case, which behaved
as I had predicted - but the behaviour changed
in Oracle 9, and I had to do some tweaking.
Turns out my test case highlighted what looks
like a but in
A bigger error in option (d) is that it leaves
open the ambiguity of whether the rows
should, or should not, be part of the answer
to the join.
Oracle's choice of join could be affected
by adding 100 rows to the table that
should be included in the join, but remain
unchanged if you add
Thanks again for all your help.
Does the ALL_ROWS hint force a preference in Join Types as well?
If so would it try to force a HASH JOIN?
From: Mladen Gogala [EMAIL PROTECTED]
Reply-To: [EMAIL PROTECTED]
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Subject: Re: Re[2]: OCP Question
PROTECTED]
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Subject: Re: Re[2]: OCP Question (Perf Tuning)
Date: Tue, 06 Jan 2004 13:24:25 -0800
Oracle doesn't but Jonathan Lewis does, in his Tutorials. I found out
about that
from Scott Gosset in 8i internals class in NYC. Seems
22 matches
Mail list logo