FROM/FROM NAMED in tdb

[Simon Helsen <sh...@ca.ibm.com>]

Hi everyone,

I was just revisiting a number of tests and I noticed that currently, we 
employ a hack for FROM and FROM NAMED queries, essentially creating an 
intermediate in-memory dataset to execute such queries. However, when I 
tried to remove this mechanism and directly execute a FROM/FROM NAMED 
query (in my particular test, it was a DESCRIBE) on a tdb dataset, I 
noticed the query's scope is not actually reduced to the graphs in the 
FROM NAMED (I am running with defaultUnionGraph=true). 

Is this a known issue or limitation? Or is there something else I am 
missing to get this work correctly?

thanks

Simon