Path: news.f.de.plusline.net!news-fra1.dfn.de!news.tele.dk!feed118.news.tele.dk!postnews.google.com!38g2000cwa.googlegroups.com!not-for-mail From: "joel garry" Newsgroups: comp.databases.oracle.server Subject: Re: What index access path does INDEX_SS yield ? Date: 4 Jan 2007 15:10:24 -0800 Organization: http://groups.google.com Lines: 45 Message-ID: <1167952224.887350.7410@38g2000cwa.googlegroups.com> References: <1167470431.542537.209020@i12g2000cwa.googlegroups.com> <1167474399.766259.93840@48g2000cwx.googlegroups.com> NNTP-Posting-Host: 75.49.200.201 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Trace: posting.google.com 1167952229 20817 127.0.0.1 (4 Jan 2007 23:10:29 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: Thu, 4 Jan 2007 23:10:29 +0000 (UTC) In-Reply-To: User-Agent: G2/1.0 X-HTTP-UserAgent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0),gzip(gfe),gzip(gfe) X-HTTP-Via: 1.0 ISA2K4-OC1 Complaints-To: groups-abuse@google.com Injection-Info: 38g2000cwa.googlegroups.com; posting-host=75.49.200.201; posting-account=YRNZ5wwAAAAg-yYjMFwy3FyWUbPiwNdq Xref: news.f.de.plusline.net comp.databases.oracle.server:191753 Jonathan Lewis wrote: > Sybrand, > > It has also been pointed out to you several times that > Oracle "hints" are not just hints in the English-language > sense of the word. In the absence of bugs, and provided > they are syntactically correct and legal, they must be > obeyed. Bugs. That kind of negates this whole paragraph, doesn't it? We can't explicate all of the internal processes that oracle goes through to determine a path, only marvel at what a trace tells us. So even though the mere fact that it is binary code makes it deterministic, the fact that we can't know all of the code adds an element of chaos. Which drives "hints" to the English meaning of the word, since we can only use heuristics to figure it out. And of course, you, Jonathan, are the best at that! Kinda makes me wonder what I'm missing if we disagree here. As many simple demonstrations have shown, chaos doesn't require many variables. http://www.physics.lsa.umich.edu/demolab/demo.asp?id=60 > > The last time you made your erroneous claim, I asked you > for an example that demonstrated your point - I am still > waiting for an answer. > I believe this thread answers that, and the answer to Spendius' original question is "yes, changing the relative volumes of data can affect the path chosen even with a hint." Somehow I wonder if we are stumbling around a strangeness with global indices here. Something like "expect sparseness from the point of view of a field within a global index so increase the cost of a skip scan." jg -- @home.com is bogus. "The Saints Are Coming... and their quarterback used to play for us." - San Diego DJ introducing a U2 song.