- From: Steve Harris <steve.harris@garlik.com>
- Date: Thu, 7 Oct 2010 21:48:52 +0100
- To: Lee Feigenbaum <lee@thefigtrees.net>
- Cc: Andy Seaborne <andy.seaborne@epimorphics.com>, Axel Polleres <axel.polleres@deri.org>, SPARQL Working Group <public-rdf-dawg@w3.org>
On 2010-10-07, at 18:53, Lee Feigenbaum wrote: > On 10/7/2010 10:24 AM, Andy Seaborne wrote: >> > On 2010-10-07, at 02:50, Axel Polleres wrote: >> > >> >> When thinking about practical examples for GROUP_CONCAT, I can >> hardly think of any where I wouldn't want to impose an order... >> >> ... thus, what's the opinions about adding another scalar parameter >> "order" which takes as parameter "DESC"|"ASC" >> >> >> >> I agree that it makes GROUP_CONCAT rather more useful to be able to >> control the order. >> >> On 07/10/10 11:18, Steve Harris wrote: >>> My feeling is that you need ORDER BY expression to make it useful. >> >> Agreed. > > I'll be a dissenting voice here -- we use GROUP_CONCAT all the time without ordering, and I'm not totally clear what use cases are addressed by ordering *based on the concatenated string values*. Could you help me understand this? To be clear. I don't think that it needs ordering to be useful - I use it without ordering in SQL all the time. I think that for /ordering/ to be useful it needs to be by expression, I have usecases for ORDER BY DESC(strlen(?name)) for e.g. > (I see how useful it would be to be able to order the values based on some other variable from the group in order to have consistent orderings across multiple GROUP_CONCAT invocations within a project list, but that's different from what I think I'm reading here...) - Steve -- Steve Harris, CTO, Garlik Limited 1-3 Halford Road, Richmond, TW10 6AW, UK +44 20 8439 8203 http://www.garlik.com/ Registered in England and Wales 535 7233 VAT # 849 0517 11 Registered office: Thames House, Portsmouth Road, Esher, Surrey, KT10 9AD
Received on Thursday, 7 October 2010 20:49:28 UTC