On 17-02-12 15:50, Randolph Neall wrote: > >Other models I didn't try yet are Object Oriented DBs and Document > Oriented DBs (XML, JSON, ...) [6]. I think DODBs are a good option, > fast for store highly hierarchical structures, but you need to write > some ugly queries if you want your data back :D > Aren't several major OpenEHR systems using XML or stored blobs on top > of a rather minimal DB schema? > The problem with ORM systems would ultimately be a rather bloated > schema and hard-wired classes to accommodate that schema. Yes, ORM > lets you automate the creation of DB data structures and classes, but, > once created, they become part of the schema and your DLLs, which is > fine until you have many hundreds of them (tables and classes defined > in autogenerated code, etc.).
That is my experience too, that is why I prefer a simple schema, for example a key(path)/value schema. I heard about good experiences with storing XML-pathes and values. The bad thing is, as Pablo said, you have to write software-logic to translate the paths/values to RM-instances and back, but the good thing is, you only need to do it once, because the kernel/RM specs are very stable. Also you need to be prepared for AQL queries and templates which use paths to retrieve information. regards Bert > Randy Neall > Veriquant, L.L.C. > > On Thu, Feb 16, 2012 at 5:26 PM, pablo pazos <pazospablo at hotmail.com > <mailto:pazospablo at hotmail.com>> wrote: > > Hi M?rcio, > > There is no standard persistence model, the persistence mechanism > is not in the standard scope. > > There are many ways of storing openEHR RM instances (archetyped > data), the only thing to take into account is that the information > to store will be highly hierarchical. > > Said that, in EHRGen [1] we use a relational model with an > Object-Relational Mapping [2] tool (GORM from Grails > Framework[3]). The advantage of that is that you have a complete > and validated RM instance persisted on the DB, and you can query > for complete objects or single data ELEMENTS. I've written ORM > tools myself [4] and the main problem is the amount of joins you > need to load a complete structure, but in my experience you never > load a complete structure for a real time interaction with the > user, and you alway can cach? some data. > > This approach is straight forward, because all you need are the > classes of the RM, and you delegate DB stuff to the ORM tool. > > Other models are viable too, like K/V [5] or EAV [6] approaches > (mentioned by Bert). This approaches are fast for saving and > loading data, the problem is that you need to have some complex > logic above that for constructing a complete RM instance on > memory, because K/V is a flat representation of a higly > hierarchical tree structure. > > Other models I didn't try yet are Object Oriented DBs and Document > Oriented DBs (XML, JSON, ...) [6]. I think DODBs are a good > option, fast for store highly hierarchical structures, but you > need to write some ugly queries if you want your data back :D > > Hope that helps. > > [1] http://code.google.com/p/open-ehr-gen-framework/ > [2] http://grails.org/ > [3] http://en.wikipedia.org/wiki/Object-relational_mapping > [4] http://code.google.com/p/yupp/ > [5] http://en.wikipedia.org/wiki/NoSQL > [6] > http://en.wikipedia.org/wiki/Entity%E2%80%93attribute%E2%80%93value_model > > > -- > Kind regards, > Ing. Pablo Pazos Guti?rrez > LinkedIn: http://uy.linkedin.com/in/pablopazosgutierrez > Blog: http://informatica-medica.blogspot.com/ > Twitter: http://twitter.com/ppazos > > ------------------------------------------------------------------------ > From: mdckoury at gmail.com <mailto:mdckoury at gmail.com> > Date: Thu, 16 Feb 2012 16:53:19 -0300 > Subject: openEHR - Persistence of Data > To: openehr-technical at openehr.org > <mailto:openehr-technical at openehr.org> > > > Hello guys, > > i'm starting a research about the persistence model of Archetype > data, that stores the information entered by the user of the system. > > I would like to know if there is a indication of the openEHR > standard for what kind of model schema should be used in DataBase, > and if there are researchs in this area. > > Thanks in advance, > > *M?rcio Costa* > B.Sc. in Computer Science @ Cin/UFPE > M.Sc. Candidate in Computer Science @ CIn/UFPE > MSN: mdckoury at gmail.com <mailto:mdckoury at gmail.com> > > > _______________________________________________ openEHR-technical > mailing list openEHR-technical at openehr.org > <mailto:openEHR-technical at openehr.org> > http://lists.chime.ucl.ac.uk/mailman/listinfo/openehr-technical > > _______________________________________________ > openEHR-technical mailing list > openEHR-technical at openehr.org <mailto:openEHR-technical at openehr.org> > http://lists.chime.ucl.ac.uk/mailman/listinfo/openehr-technical > > > > > _______________________________________________ > openEHR-technical mailing list > openEHR-technical at openehr.org > http://lists.chime.ucl.ac.uk/mailman/listinfo/openehr-technical -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openehr.org/mailman/private/openehr-technical_lists.openehr.org/attachments/20120217/9009309a/attachment.html>
