Hi Wahed, I spent some time looking for something like you mention, and in terms of relational DBMS, I found only this: https://github.com/renaud/uima_sql but after working on it for a while and talking with the original author, we concluded that, following Richard's sugestion, this was generally a Bad Idea.
You're much better off using some other type of DBMS. Personally, I would look into a document database that can query the XMI representation of the CASes directly along with the type system specification if you need to retain some kind of type-checking. If you decide to go that route, I would suggest BaseX: http://basex.org/, as it's what I've used locally for a 3gb corpus. Just point BaseX to a dir with the XMI files in it and you should be up and running in seconds. Then you can write all the plumbing code you need in xquery, which is much much nicer than SQL. Cheers, jta On Wed, Jan 25, 2017 at 11:55 AM <hem...@em.uni-frankfurt.de> wrote: > Hi, > is there any way to store processed CAS in a database? Does anyone > have experience with that? Which database would you recommend (SQL, > NoSQL, Graphdb,..) considering that the typesystem could be changed. > How would one map the TypeSystem to a DB? It also should be able to > manage big amout of data (TB or more). > > Thanks in advance > > -Wahed > > -- > A. Wahed Hemati > Text-Technology Lab > Fakultät für Informatik und Mathematik > Johann Wolfgang Goethe-Universität Frankfurt am Main > Senckenberganlage 31, Raum 401c > 60325 Frankfurt am Main > Postfach: 154 > Tel: +49 69-798-28925 <+49%2069%2079828925> > Email: hem...@em.uni-frankfurt.de > Web: http://www.hucompute.org/ > > -- sent from a phone. please excuse terseness and tpyos. enviado desde un teléfono. por favor disculpe la parquedad y los erroers.
