it seems to me, in your case the sequence is not part the actual data
source. If that is true, it is e.g. only a temporary condition during
runtime of an application and has nothing with data or its encoding.
Hence, if it's really too time consuming to do the sorting on demand,
I'd store sorted pointers to the data source apart from it. This
collection could include all necessary kinds of sorting without
increasing the amount of data too much, but has to be synchronized with
the underlying data when changes occur for some reason.
On 30.05.2011 15:17, Torsten Schassan wrote:
> I have to sort <idno>s according to rules that seem to be too
> complicated to be used during db lookups. Would it be reasonable to
> introduce @sortKey for <idno>? How do others cope with the problem?
> Best, Torsten
Berlin-Brandenburgische Akademie der Wissenschaften
eMail [log in to unmask]
phone +49 (0)30 20370 285