Pickling prepared SPARQL queries

[n...@mayacs.com]

Hi,

I am trying to optimize SPARQL query execution by preparing a query (as described in https://rdflib.readthedocs.org/en/latest/intro_to_sparql.html) then storing the resulting object in a memcache instance. This fails when the stored object is unpickled (traceback below). I also tried storing the the parsed query object, but this fails when the object is pickled. In borh cases, I suspect the issues is that some of the constituent objects implement __getitem__() or __setitem__() in some way that the pickler/unpickler does not like.

Has anyone seen this problem before, or can you think of a simple workaround?

Cheers,

Niels