Re: Time series database misinformation from Achmal B Chaudhrii
>>>>> "cj/3" == Colin James <cjames3@ix.netcom.com> writes: cj/3> In <Pine.SOL.3.91.960419130254.24047A-100000@swindon> Akmal B Chaudhri cj/3> <ad533@city.ac.uk> writes: >> One good source is the Illustra web site. They have developed Data-Blades >> specifically to handle this type of data. >> >> See: >> >> http://www.illustra.com/ >> mailto:info@illustra.com cj/3> Illustra is an Informix "company", but Achmal is way off the mark with cj/3> the DataBlade stuff and time series. His name is Akmal. cj/3> DataBlade modules are meant to be simpler defined-functions functions cj/3> for data types concocted by the user which return values. cj/3> So what. cj/3> Time series are a whole different species. cj/3> Answer the question next time Achmal, rather than throwing up cj/3> information, then dancing. Akmal DID answer the question. Time series data are simply data that are inherently ordered. They are typically measurements of some kind, (financial, process, etc), that are taken periodically through time. The key point is that they are collected in order, and it almost always makes sense to process them only in that same order, and the order is an important part of the data. RDBMS are horrible at this, because their very definition is un-ordered sets. Thus, you end up spending huge amounts of time searching and sorting on artificial numerical indexes, just to continuously rebuild the order you started with. ODBMS are much better, because they allow the direct representation of ordered data structures, such as arrays of objects. This is very similar to the issue of spacial (geographical) data, where (as CJ properly pointed out some time ago), RDBMS perform terribly. And, Illustra DataBlades work for geographical data too, (though probably not as well as an ODBMS). Don Don Erway derway@ndc.com NDC Systems 818-939-3847 5413 N. Irwindale Ave Fax:939-3870 Irwindale, CA, 91706 Disclaimer: These are my own personal opinions, and not those of my company, or anyone else.