0

I'm thinking that to create two dataset, one is in memory, which frequently changed, another is tdb, which is read-only, can be created as one dataset, but in the back, which in memory remain in memory, and which in tdb remain in tdb. If not, can two tdb dataset be union as one dataset?

Gao
  • 912
  • 6
  • 16
  • You can put graphs from different datasets into a new dataset. – AndyS May 07 '19 at 15:29
  • @AndyS sorry to trouble, But if I made one dataset in memory, another in tdb, whether union them in memory or in tdb? – Gao May 08 '19 at 06:28
  • The general dataset, with graphs from each of the other datasets (limitation - each named graph in the general dataset can only come from one other dataset) leaves the graph in its original storage and directs request to the right place. – AndyS May 08 '19 at 10:33
  • @AndyS thanks, I have done a little research for this, for now I can make a default graph which union two different backend graph(one is tdb, one is in memory). But I still have a problem, when I query the default query, how can fuseki know when to use memory db and when to use tdb? – Gao May 13 '19 at 10:44
  • I have taken a peek about jena's code, I found in "MultiUnion" class, it delegate the query to subGraph which it contains. I don't know whether I am correct. – Gao May 13 '19 at 11:08
  • yes - "multiunion" itself has no state - it accesses the subgraphs and makes sure the triples are not returned more than once. – AndyS May 13 '19 at 17:01

0 Answers0