> Here's the context: We have several hundred thousand time series, but > they will, for the foreseeable future, remain in their present > repository (the Fame time series database application.) > > However, there are a few hundred series which our staff use > regularly, and which they need to query, graph, plug into reports, > save as PDFs, import into spreadsheets, etc. By and large, the Cocoon > framework seems like a promising way of providing these "multiple > views" of the data. > > At this early stage in our planning, we're considering a mechanism > whereby these few hundred series are dumped from Fame into XML, one > file per time series. We explicitly don't want to dump the data into > some intermediate format (e.g, Oracle or whatever), because Fame will > remain the core database.
I suppose that you can't hit "Fame" (whatever that might be) directly? Or that you are afraid of the load that hitting "Fame" directly would put on the system? I don't really see why dumping the data into Oracle or whatever is any different from creating an intermediate XML file? They are both data stores... You need to determine what the functional requirements are in terms of performance and querying and let that determine the technical architecture and not just pick operating system managed XML files arbitrarily. However, either way Cocoon still may be a good fit; it can handle output from an relational database (with JDBC) almost as easily as handling an XML file directly and if you have to create the XML from scratch it probably makes more sense to do that with Cocoon instead of using some other external tool to do the job. --------------------------------------------------------------------- Please check that your question has not already been answered in the FAQ before posting. <http://xml.apache.org/cocoon/faq/index.html> To unsubscribe, e-mail: <[EMAIL PROTECTED]> For additional commands, e-mail: <[EMAIL PROTECTED]>
