- From: Richard Cyganiak <richard@cyganiak.de>
- Date: Wed, 1 Jun 2011 22:38:36 +0100
- To: Eric Prud'hommeaux <eric@w3.org>
- Cc: David McNeil <dmcneil@revelytix.com>, W3C RDB2RDF <public-rdb2rdf-wg@w3.org>
On 1 Jun 2011, at 22:16, Eric Prud'hommeaux wrote: > I was disputing the opacity of the query. A user would want to identify the language extensions used in particular queries if one query used, for instance, geo extensions and another some simple statistics, and you wanted to give the R2RML engine the best chance of optimizing the queries that it understood. This makes absolutely no sense to me. To optimize a query, an engine has to parse a query. If the query doesn't parse, then the engine doesn't understand it and has to treat it as opaque. And in reality, your engine will be able to optimize queries that only involve inner and left joins, but won't be able to optimize queries that include subqueries and aggregates. This flagging business is an additional burden on the user (and on the WG, if it means minting and maintaining IRIs for SQL dialects), with the only benefit that an R2RML engine can flag certain syntax errors. Best, Richard
Received on Wednesday, 1 June 2011 21:39:05 UTC