Path: newssvr20.news.prodigy.com!newsmst01a.news.prodigy.com!prodigy.com!news.glorb.com!postnews1.google.com!not-for-mail From: gnuoytr@rcn.com (robert) Newsgroups: comp.databases.theory Subject: Re: Teach SELECT DISTINCT first! Date: 27 Apr 2004 13:41:26 -0700 Organization: http://groups.google.com Lines: 28 Message-ID: References: NNTP-Posting-Host: 20.137.30.50 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1083098486 16894 127.0.0.1 (27 Apr 2004 20:41:26 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: Tue, 27 Apr 2004 20:41:26 +0000 (UTC) Xref: newssvr20.news.prodigy.com comp.databases.theory:26022 "Laconic2" wrote in message news:... > Thanks. I didn't know that about Date pontificating. I don't read that > much Date. > > That's really what I was driving at, a little while ago, when I said that > SQL had made an unfortunate choice in the meaning of "SELECT" > > My thought is that, if SQL had originally made "SELECT" default to "SELECT > DISTINCT" instead of "SELECT ALL", there would be less misunderstanding of > the RDM than there is. > > But I was declared "orthogonal". > > Note that, when it came to "UNION" SQL defaulted to "UNION DISTINCT", and > if you want "UNION ALL", you have to say "UNION ALL". > > I claim, without proof, that a really good optimizer can tell, by using > rules of logic, and perhaps more metadata than is now stored, when "SELECT > ALL" and "SELECT DISTINCT" will produce the same result. If so, it can > skip a step, and thereby speed things up.\ i'm way too lazy to type all this in (or take credit for it). http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&threadm=c0d87ec0.0202240850.428d7b88%40posting.google.com&rnum=12&prev=/groups%3Fq%3D%252B%2522select%2Bdistinct%2522%2B%252Bdate%2B%2Bgroup:comp.databases.theory%26start%3D10%26hl%3Den%26lr%3D%26ie%3DUTF-8%26oe%3DUTF-8%26group%3Dcomp.databases.theory%26selm%3Dc0d87ec0.0202240850.428d7b88%2540posting.google.com%26rnum%3D12 if one wishes not to read it: Celko cites Codd and Date in arguing against dups. i concur.