Path: csiph.com!x330-a1.tempe.blueboxinc.net!usenet.pasdenom.info!gegeweb.org!de-l.enfer-du-nord.net!feeder2.enfer-du-nord.net!weretis.net!feeder1.news.weretis.net!news.solani.org!.POSTED!not-for-mail From: Mladen Gogala Newsgroups: comp.databases.postgresql Subject: Re: Is PostgreSQL good? Date: Thu, 30 Jun 2011 22:42:19 +0000 (UTC) Organization: solani.org Lines: 42 Message-ID: References: <1308640710.210659@proxy.dienste.wien.at> <1308738638.232318@proxy.dienste.wien.at> <1308911383.819034@proxy.dienste.wien.at> <87vcvnxqsj.fsf@araminta.anjou.terraraq.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-Trace: solani.org 1309473739 17718 eJwNxsEBwCAIA8CVREik4wDC/iPYex2UwjpG0DAY/1dDrtgr4sBCBd55+7Zn5mbl15AEldznAR8rES8= (30 Jun 2011 22:42:19 GMT) X-Complaints-To: abuse@news.solani.org NNTP-Posting-Date: Thu, 30 Jun 2011 22:42:19 +0000 (UTC) User-Agent: Pan/0.133 (House of Butterflies) X-User-ID: eJwFwYEBwDAEBMCVKE+NI+T3HyF3sNCY9EA4CH5HxheUzLoL9+3W7RAUTPs6ey4LBz/Wkg8lExGU Cancel-Lock: sha1:tXAf8dlM73Ms02Pqg6YPd7QRzKc= X-NNTP-Posting-Host: eJwNyMEBwEAEBMCWHHZRzkXov4RkngPjYYcTdCyWU0OX1RiYQO7mv5NP8/pooDQp/TqMFWc+GZIQuw== Xref: x330-a1.tempe.blueboxinc.net comp.databases.postgresql:172 On Thu, 30 Jun 2011 13:21:17 -0700, Don Y wrote: > Gee, someone expresses an interest in understanding why you have such a > hard-on regarding "hints" and you dismiss it as "irrelevant". What I consider irrelevant is the analogy with the C language. Furthermore, this debate is, at least in my humble opinion, relevant to DBA's, project managers and people who will write SQL against a Postgres database. I haven't envisioned this debate as a talk show with the viewers taking sides, Oprah already covers that market. As far as me having hard-on for hints, I needed them in one of my projects. Hints are a welcome option, offered by virtually all other RDBMS systems, yet there seems to be a problem with the designers of Postgres. I have yet to see a logical argument against providing that option and analogies with the C programming language don't seem to provide it, either. So, those analogies are irrelevant, at least in my point of view. You don't seem to be a database professional either, I am quite confused by the zeal with which you're trying to participate in this debate. Once again, this debate is about a specific feature of Postgres, or the lack of it. FYI, hints are a method to force the optimizer to do what the programmer wants, rather than what the RDBMS thinks is right. Here are two implementations of query hints: http://dev.mysql.com/doc/refman/5.5/en/index-hints.html http://msdn.microsoft.com/en-us/library/ms181714.aspx There are, of course, other implementations, in DB2 and Oracle. As a matter of fact, the only other relatively well known database that didn't have hints was Firebird. Fortunately, its designers have seen the light and the hints are planned for the version 3.0: http://www.slideshare.net/mindthebird/initial-review-of-firebird-3 The argument is whether the users should have power to override the optimizer decisions. I haven't seen a single coherent argument why they shouldn't have that option and I definitely did have a need for hints. -- http://mgogala.byethost5.com