On 15 Nov 2010, at 3:37 PM, Mikael Syska wrote: > I think we have talked about it before, a new database design, to > optimize speed when searching and avoid duplicate data. Right now most > quries are full table scans and most data are stored a strings.
The format of the data logs is better stored in a NOSQL DB, and the queried via a map reduce implementation. I experiemented with both couchDB and mongoDB mongo seems to be the best suited unfortunately their map reduce implementation is not fast enough at the moment, but they are working on fixing this. So for now any SQL based implementation would just be a workaround using possible procedures and triggers that is why its not on the roadmap. > > Is this dropped or just forgot in the roadmap ? If anyone has an approach to this that does not involve aggregation using triggers or procedures then lets discuss it and we can implement that. > > Think you said it was a feature for 1.0 ... -- Baruwa - www.baruwa.org _______________________________________________ Baruwa mailing list [email protected] http://lists.baruwa.org/mailman/listinfo/baruwa
