This discussion is archived
1 7 8 9 10 11 Previous Next 163 Replies Latest reply: Oct 22, 2009 3:23 PM by Hoek Go to original post RSS
  • 135. Re: Index rebuild
    706417 Explorer
    Currently Being Moderated
    Hi Don,

    As you say, "Browning" is a useful term. An index is tree-like, after all.

    And the Great Tom Kyte used that expression on Ask Tom.

    In the real world, loads of quick, but meaningful, phrases are used - and, miraculously, the world keeps turning and everyone knows what those expressions mean.


    Regards - Don Lewis
  • 136. Re: Index rebuild
    Hoek Guru
    Currently Being Moderated
    And the Great Tom Kyte used that expression on Ask Tom.
    Well, he 'used' it on AskTom, yes, but it had nothing to do with index rebuilds at all, but with "the time to perform instance recovery for the failed node", see: http://asktom.oracle.com/pls/asktom/f?p=100:11:0::::P11_QUESTION_ID:18305250155354#64198627160806
    Feel free to post some proof if I'm wrong, I'd be happy to know.
    I'm wondering anyway why you're not posting some proof (a simple link) when you're making statements about other persons like that.
  • 137. Re: Index rebuild
    706417 Explorer
    Currently Being Moderated
    Hi hoek,

    I typed this into Google: "*tom kyte index rebuild browning*"
    and got this: http://asktom.oracle.com/pls/asktom/f?p=100:11:956520231236179::::P11_QUESTION_ID:6601312252730

    Tom:

    I have a few Questions Re: Index characterisitcs. I'd grealty appreciate if you could please
    explain with the help of some samples.

    How can I determine if my index is
    +1) "Lob sided" and what does it mean ?+
    +2) Index browning - what is it and how to detect it?+
    +3) undergoing block splitting ?+
    +4) sequencially increasing in its key value - and hence I can re-create it with a low pctfree+
    setting.

    How can I determine the overhead of a bulk update/insert on the index maintenance (i.e block
    splitting at leaf level and branch level etc...)?

    Thanks




    Followup   May 2, 2003 - 12pm Central time zone:

    +1) think you mean "lop sided" and for that I would ask you to ask the person that said it what they+
    mean and how to get one ;)
    they really don't get that way.

    *+2) a high percentage of deleted leaf rows to leaf rows (think of leaf rows like leaves on a tree +*
    dead leafs (deleted row entries) are "brown".  If you have lots of dead leaves, the tree looks*
    brown)*


    I know, I know. What can I say? It's a gift, perhaps even magic. I just hope that I can continue to use it for the betterment of all.


    Love - Don Lewis
  • 138. Re: Index rebuild
    Hoek Guru
    Currently Being Moderated
    I typed this into Google: "tom kyte index rebuild browning"
    ¿Now that's 'ironic'/strange?

    If one does a search on 'browning' on http://asktom.oracle.com, your googled link doesn't even show up at all in the results...
    It seems that since AskTom has moved, some search results tend to differ since then. There are other 'classic threads' not showing up anymore, I already found out.
    I know, I know. What can I say? It's a gift, perhaps even magic. I just hope that I can continue to use it for the betterment of all.
    Who knows, Don Lewis...only Google and time can tell, or perhaps a person called Jonathan Burleson ;)
  • 139. Re: Index rebuild
    Jonathan Lewis Oracle ACE Director
    Currently Being Moderated
    burleson wrote:

    And yes, I'm only a DBA, but I'm a lot smarter than you are, and I have a better memory too . . .
    There are several articles on my website ( http://www.jlcomp.demon.co.uk ) demonstrating this fact
    Right, sure there are . . .
    See the article on my website ( http://www.jlcomp.demon.co.uk ) about large keys and excessive leaf block splits
    If that's true, why are you reluctant to give us an executive summary?
    Personally I think the expression "index browning" is a silly one
    As a teacher, I think it's brilliant.
    Your memory must be giving you trouble - only six lines down the page and you've forgotten that you're a DBA.

    Regards
    Jonathan Lewis
    http://jonathanlewis.wordpress.com
    http://www.jlcomp.demon.co.uk

    To post code, statspack/AWR report, execution plans or trace files, start and end the section with the tag {noformat}
    {noformat} (lowercase, curly brackets, no spaces) so that the text appears in 
    fixed format
    .
    
    +"Ridicule is the only weapon that can be used against unintelligible propositions.+
    +Ideas must be distinct before reason can act upon them."+
    Thomas Jefferson
    
    
    P.S.  If you really intended to honour Justin's request, you wouldn't keep going back and editing the posts you've already made in this thread.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        
  • 140. Re: Index rebuild
    706417 Explorer
    Currently Being Moderated
    DKB has memory issues? You mean like the one you have?

    Only [very recently|http://forums.oracle.com/forums/thread.jspa?threadID=972112&tstart=0], you called for this thread to be locked (remember?), yet you still post on it!

    Quick, nurse, the screens!

    Love - Don Lewis
  • 141. Re: Index rebuild
    Jonathan Lewis Oracle ACE Director
    Currently Being Moderated
    burleson wrote:
    ... responds with something out of left field, like his critique of an unrelated article . . .
    Dear Mr. Burleson,

    If [+*my critique*+|http://forums.oracle.com/forums/thread.jspa?threadID=963129&tstart=0] is of an unrelated article, why did you introduce it to the thread in the first place: Re: Index rebuild

    Regards
    Jonathan Lewis
    http://jonathanlewis.wordpress.com
    http://www.jlcomp.demon.co.uk


    To post code, statspack/AWR report, execution plans or trace files, start and end the section with the tag {noformat}
    {noformat} (lowercase, curly brackets, no spaces) so that the text appears in 
    fixed format
    .
    
    
    Extraordinary claims require extraordinary evidence
    Carl Sagan                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        
  • 142. Re: Index rebuild
    Jonathan Lewis Oracle ACE Director
    Currently Being Moderated
    burleson wrote:

    Please don't assume that I wrote everything, I only contribnute a small fraction of the knowledge . . .
    Since an index is by definition ordered, then rebuilding it will never change the clustering factor.
    That's not how I read it, but I agree that sentence might be misunderstood by some people.

    That's a book excerpt by Mike Ault that I purchased rights to, but I fixed it to make this clear that he is referring to rebuilding the table, not the index.
    How sweet, now that Mike Ault doesn't work for you, you don't mind blaming the errors on him. So here's a reference to an item you posted in March 2007 which shows that you thought that rebuildling the index would change the clustering_factor: Re: rebuilding indexes


    Regards
    Jonathan Lewis
    http://jonathanlewis.wordpress.com
    http://www.jlcomp.demon.co.uk


    To post code, statspack/AWR report, execution plans or trace files, start and end the section with the tag {noformat}
    {noformat} (lowercase, curly brackets, no spaces) so that the text appears in 
    fixed format
    .
    
    
    "Your manuscript is both good and original; but the part that is 
    good is not original and the part that is original is not good."
    Samuel Johnson (attributed)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    
  • 143. Re: Index rebuild
    706417 Explorer
    Currently Being Moderated
    Which bit says that? It's a groggy morning for me, so I may be missing the glaringly obvious.


    Regards - Don Lewis
  • 144. Re: Index rebuild
    6363 Guru
    Currently Being Moderated
    burleson wrote:
    Hi Pointless,
    Possibly because when previous indicators have been identified they have been proven to be incorrect.
    You still believe that "one negative test case proves something wrong" nonsense?

    Test cases are painfully easy to rig!

    I can write one to disprove anything you want, it's easy, try it!
    Ok since you ask, publish a test case that shows an index rebuild changing the clustering factor
    BTW, re-read the credits on that article. As you see, I did not write that!
    Correct you published the book by an author who worked for your consulting firm recommending monitoring indexes for rebuilding by examining the clustering factor.

    You then republished it on a commercial consulting site under the heading Oracle Tips by Burleson. I never said you wrote it but under the circumstances I thought you may have read it and had perhaps tried to understand the tips you promote.
  • 145. Re: Index rebuild
    6363 Guru
    Currently Being Moderated
    Don Lewis wrote:

    Has nobody ever been involved in a User Acceptance Test, ironed out the problems, re-run with 100% success, and then seen it crash and burn in the production release
    Yes, but it doesn't mean unit tests, regression tests etc are useless or meaningless, just that they have to carefully constructed to account for all variables and that it is possible to not take something into consideration that should have been. Result is a bug. In software development what typically happens is that once the bug is identified the tests are updated and refined to correctly test that scenario and the tests and software quality improve.

    What does not happen for very good reasons is the development team throw their hands up and say "This is too complicated, just like magic, lets not bother with any testing since it is all meaningless and let the customers just run it and hire a self proclaimed expert to apply a mystic vision to the problem". Which appears to be what you and the self proclaimed expert would want us to believe.

    >
    To recap: Oracle RDBMS is not science.
    Yes it is. Computer programs are the results of computer science that is studied at universities and has theories that can be tested just like medicine. I suppose you don't consider a medicine a science since you can give the same treatment to ten patients and nine will recover and one won't. The fact that both are complicated and you don't fully understand them doesn't make them not sciences.
    It is a big bunch of software. It has bugs and stupidities in it.
    Yes it is complicated just like medicine, I suppose since the results of medical treatment are unpredictable the US system could be reformed simply by stopping all spending on developing new drugs and running hospitals and call in some witch doctors instead.
    Someone else's test cases are of interest, sure, but do not have meaningful relevance to my system. Only my test cases have meaningful relevance to my system. Demonstrating anything on a pared-down Oracle database on a laptop and suggesting that a big fat multi-terabyte, SANned-up, Infinibanded, RAC-enabled, you name it system, with its attendant Oracle-supplied params fixes and tweaks is likely to behave in an identical manner is a bit unconvincing.
    I would expect them to implement the same B*tree index structure though, wouldn't you?
  • 146. Re: Index rebuild
    706417 Explorer
    Currently Being Moderated
    +"Computer programs are the results of computer science that is studied at universities and has theories that can be tested just like medicine"+

    Then you won't mind my one Test Case blowing that twaddle to atoms? Good. Here goes, then:
    I wrote a program today. A simple little thing, but lovely and perfectly functional, nonetheless. I have no degree in any science. So, no Computer Science was used to create that program. Thus your Law of Nature-like fiat: "Computer programs are the results of computer science that is studied at universities and has theories that can be tested just like medicine" bit of "I can pronounce Kierkegaard" philosophy trips, stumbles and scrapes its knees, only to later die horribly of complications relating to it being a positive flake of dandruff-like joke.


    To recap: Oracle RDBMS is not science.
    +"Yes it is"+

    Oh no it isn't!

    Which reminds me, I must buy the panto tickets soon.

    Anyway, Medicine is a separate entity to Science, hence terms such as Science-Based Medicine, as opposed to, say, accupuncture, which is still (and I bet forever) balancing on the edge of real/fake medicine, even though it's ( [in the UK|http://www.nhs.uk/Conditions/Acupuncture/Pages/Who-can-use-it.aspx] , for instance) a recognized discipline of medicine.

    The Sciences can be used in Medicine, but Medicine isn't Science. Oh! How those without a science degree love to twist a few years' experience or a degree in Arran Knitting into a Double First from Cambridge!


    Love - Don Lewis
  • 147. Re: Index rebuild
    6363 Guru
    Currently Being Moderated
    Don Lewis wrote:
    +"Computer programs are the results of computer science that is studied at universities and has theories that can be tested just like medicine"+

    Then you won't mind my one Test Case blowing that twaddle to atoms? Good. Here goes, then:
    I wrote a program today. A simple little thing, but lovely and perfectly functional, nonetheless. I have no degree in any science. So, no Computer Science was used to create that program.
    And you wrote your own compiler and built your own computer also? Very impressive.

    Or maybe you didn't, and computer science was very much present in your simple little program.
    Thus your Law of Nature-like fiat: "Computer programs are the results of computer science that is studied at universities and has theories that can be tested just like medicine" bit of "I can pronounce Kierkegaard" philosophy trips, stumbles and scrapes its knees, only to later die horribly of complications relating to it being a positive flake of dandruff-like joke.
    If you want to believe software has nothing to do with computer science, that there is no point in testing software functionality for quality, that all software behaves randomly and can only be improved or configured by someone who tells you that they are an expert and you just believe them then the good news is that [you are not alone|http://www.alaska.net/~clund/e_djublonskopf/Flatearthsociety.htm]. These are public forums though so not everyone is as gullible as you strive to be.

    >
    Anyway, Medicine is a separate entity to Science,
    Because you say so. And you use computers that have been created without involvement of any computer scientists.

    Oh look, a web page listing the basic sciences of medicine

    http://en.wikipedia.org/wiki/Medicine#Basic_sciences
    The Sciences can be used in Medicine, but Medicine isn't Science. Oh! How those without a science degree love to twist a few years' experience or a degree in Arran Knitting into a Double First from Cambridge!
    So just like software? Do you know its kind of hard to follow what you are going on about much of the time. I am not sure since you don't believe medicine to be a science and you don't believe software to be a science that you apply the same attitude to both. If something were wrong with you, would you say "Don't give me any of your meaningless tests, just try sticking some pins in me and bring on the chicken blood."
  • 148. Re: Index rebuild
    user503699 Expert
    Currently Being Moderated
    Don Lewis wrote:
    Which bit says that? It's a groggy morning for me, so I may be missing the glaringly obvious.
    Rebuilding the index removes the space wastage (making fast-full index scans faster, and improving the spped of >index range scans). It also places the index keys in the same order as the table rows.
    This para, and the statement in bold in particular, indirectly suggests that clustering factor of the index can be improved by rebuilding the index (unless Don says that it is a typo :) or I am mistaken that having index keys in the same order as the table rows results in better clustering factor. :) )
  • 149. Re: Index rebuild
    706417 Explorer
    Currently Being Moderated
    +"And you wrote your own compiler and built your own computer also? Very impressive.+
    +Or maybe you didn't, and computer science was very much present in your simple little program."+

    I see. So merely by using something that some other has created using the tools of science makes me, by proxy, so to speak, a scientist. Let's put that to the test:
    The television was created using a lot of science.
    I watch TV.
    How much scientific knowledge do I require to do that? Er, none. Oh dear. You've bombed-out again.


    +"If you want to believe software has nothing to do with computer science, that ... etc."+

    Not so much my wanting to believe it, rather that it's a truth and can't be avoided. I liked your Flat Earth link - very imaginative.


    +"Oh look, a web page listing the basic sciences of medicine"+

    Er, I said "The Sciences can be used in Medicine, but Medicine isn't Science." And your Wiki link simply says "Here is a list of sciences that are related to Medicine".


    +"Do you know its kind of hard to follow what you are going on about much of the time"+

    I didn't know, but I had a strong suspicion (I have read your posts before).


    +"If something were wrong with you, would you say "Don't give me any of your meaningless tests, just try sticking some pins in me and bring on the chicken blood.""+

    No. I'd say "Give me that medicine that has, so far, been shown to be reasonably safe, even though you tested it on a relatively small number of humans". The alternative is to wish away disease, which isn't something I have much of faith in. Here's a nice, big list of medicines (every one of 'em a scientifically proven winner - if you believe that testing stuff and it not wiping-out the subjects in the first few trials is science - like you probably do) that have been withdrawn:
    http://en.wikipedia.org/wiki/List_of_withdrawn_drugs


    I love you Pointless and always have done.

    Don Lewis x

Legend

  • Correct Answers - 10 points
  • Helpful Answers - 5 points