- From: Richard Cyganiak <richard@cyganiak.de>
- Date: Fri, 27 Jan 2012 14:25:57 +0000
- To: Eric Prud'hommeaux <eric@w3.org>
- Cc: Ted Thibodeau Jr <tthibodeau@openlinksw.com>, Juan Sequeda <juanfederico@gmail.com>, Ivan Herman <ivan@w3.org>, Alexandre Bertails <bertails@w3.org>, W3C RDB2RDF <public-rdb2rdf-wg@w3.org>
Hi Eric, On 26 Jan 2012, at 20:51, Eric Prud'hommeaux wrote: > I would expect, however, that many readers of the document will have > MySQL or Postgresql and have not even seen quotes around identifiers > before. In fact, MySQL doesn't even parse ""s around identifiers (even > in ANSI compatability mode, per my tests). In MySQL it's `Addresses`. This notation is all over PhpMyAdmin, so MySQL users tend to be familiar with it and certainly have seen quotes around identifiers. > Quoting consistency is low: > <http://www.alberton.info/dbms_identifiers_and_case_sensitivity.html>. > Unfortunately, that article's advice: > [[ > The safest choice if you care about portability and peace of mind is > not to quote the identifiers when you create your tables/fields and > when you run your queries. > ]] > doesn't work for us as the unquoted direct graph would look like > (according to this test -- I haven't confirmed beyond playing with > MySQL's ANSI mode and seeing no change in behavior): > > Oracle: <PEOPLE/ID-7> <PEOPLE#FNAME> "Bob" . > PostgreSQL: <people/id-7> <people#fname> "Bob" . > MySQL/Windows: <people/ID-7> <people#fname> "Bob" . > MySQL/Unix: <People/ID-7> <People#fname> "Bob" . > SQLite: <People/ID-7> <People#fname> "Bob" . > SQL Server: <People/ID-7> <People#fname> "Bob" . Standard SQL 2008: <PEOPLE/ID-7> <PEOPLE#FNAME> "Bob". You'll either have to use that, or quote all identifiers in the DM spec. Richard
Received on Friday, 27 January 2012 14:26:28 UTC