Am Mittwoch, den 10.01.2007, 17:53 +0000 schrieb robert burrell donkin: > On 1/10/07, Joachim Draeger <[EMAIL PROTECTED]> wrote:
> > For instance fetching a small subset of the headers of many messages has > > to be done in an optimal way. (e.g. in one SQL query) > > not just a single query but an efficient single query. it's not good > enough to sticking everything in one huge table and hoping for the > best. It's not? ;-) > clients are typically interested only in a subset of the headers Right. For current implementation: I already split the headers up in single rows for this purpose. (And for efficient searching) Unfortunately the current FetchCommand implementation always examines the whole MimeMessage. I'm going to fix this first and than provide a way to query a header subset. > think about this from the perspective of meta-data. when using a RDBMS > for storage, it's not efficient for all meta-data to be equal. some is > much more important that others. this should be reflected in the > database structure. Could you give an example? > IMHO this optimization is best done in the storage layer Right, but you have to provide methods in the API for such requests. That's why there are so many more or less protocol aware methods. A possibility for requesting a subset of headers is still missing. Joachim --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
