Hi Adriano, Am 21.11.2011 20:04, schrieb Adriano dos Santos Fernandes: > On 21/11/2011 16:45, Frank Ingermann wrote: >> So this would mean option b) is correct: >> >> b) [ ] window functions are not supposed to work in the "child select" >> part of recursive CTEs >> >> - since Window functions belong to<aggregate function> |<grouping >> operation> ? > > They are not part of these clause, so it should be supported, and PgSQL > supports it.
hmm... the fact that PG "supports" it (=doesn't throw an error) doesn't necessarily make it a good choice for doing what i was trying to do - (i learned -the hard way, over the years- that results you GET may coincide with what you EXPECT, and that can fool you into thinking your assumptions were right - but that does not VERIFY that the way you're trying to do things is the RIGHT way (nor that your assumptions were right.) It's difficult. >> (I doubt that would be a safe way to do that "implicit" ordering in >> a consistent way, anyway - like what if you use TWO window functions?) >> > Yes, it's unspecified. OK, i can live with that. (and could you please stay with that, otherwise i'd have to revise my slides for the upcoming conference for the 47th time or so! ;-) ) cheers, Frank ------------------------------------------------------------------------------ All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel