I just pushed some improvements here. https://github.com/plt/racket/commit/744fde1fa85aa52cd9a969d9deffa726ca94af37
Jay On Sat, Dec 21, 2013 at 12:58 AM, Evan Donahue <[email protected]> wrote: > That would be a welcome change for me as well. It would be nice to be able > to interact with the theory more readily from an embedding, interactive > application. > > Thanks, > Evan > > > On Fri, Aug 16, 2013 at 11:29 AM, Jay McCarthy <[email protected]> > wrote: >> >> You could eval the datalog macro to do it at runtime. >> >> I could also expose the underlying protocol that ! and ? use, which IS >> based on runtime symbols. That would be more like exposing a database >> engine API than a Datalog API though. >> >> Jay >> >> On Fri, Aug 16, 2013 at 11:04 AM, Mikko Tiihonen >> <[email protected]> wrote: >> > Thank you for the clarification! >> > >> > Do I understand right that since the tables do not have any binding, >> > there is no way to parameterize them in runtime? The macro approach works >> > if >> > the table is known in compile-time, but I am at loss to think of a way of >> > providing the table to the query in a variable that gets its value during >> > runtime. >> > >> > Actually, since I'm doing just some preliminary exploration, this was a >> > valuable and timely piece of advice and I will return to the drawing board >> > and take this into consideration. >> > >> > Have a nice weekend! >> > >> > Best, >> > >> > Mikko T. >> > >> > Jay McCarthy kirjoitti 16.8.2013 kello 19:36: >> > >> >> The table position is not an expression position because tables are >> >> not values and do not have any binding. In particular, they are not >> >> symbols as you assume in your example. >> >> >> >> I could imagine trying to supporting something like your example, but >> >> it would be backwards incompatible and unlike any other Datalog system >> >> as far as I can tell. Do you have any opinions about this? >> >> >> >> You can do something similar with a simple macro: >> >> >> >> (define-syntax-rule (param-query table) >> >> (datalog family >> >> (? (table 'joseph2 X)))) >> >> >> >> (param-query parent) >> >> >> >> On Fri, Aug 16, 2013 at 5:08 AM, Mikko Tiihonen >> >> <[email protected]> wrote: >> >>> Hi! >> >>> >> >>> I've been trying to parameterize Datalog queries and have stumbled >> >>> into a problem with the tables in queries. As an example (from Datalog >> >>> documentation): >> >>> >> >>> #lang racket >> >>> (require datalog) >> >>> (define family (make-theory)) >> >>> (datalog family >> >>> (! (parent joseph2 joseph1)) >> >>> (! (parent joseph2 lucy)) >> >>> (! (parent joseph3 joseph2))) >> >>> >> >>> -now if I run the query as shown in the documentation >> >>> >> >>> (let ([x 'joseph2]) >> >>> (datalog family >> >>> (? (parent x X)))) >> >>> >> >>> the answer is as expected: >> >>> >> >>> '(#hasheq((X . joseph1)) #hasheq((X . lucy))) >> >>> >> >>> However, if I parameterize the table: >> >>> >> >>> (let ([x 'parent]) >> >>> (datalog family >> >>> (? (x 'joseph2 X)))) >> >>> >> >>> the query fails: >> >>> >> >>> '() >> >>> >> >>> Are there any workarounds for this? I am grateful for any >> >>> suggestions... >> >>> >> >>> Best regards, >> >>> >> >>> Mikko Tiihonen >> >>> >> >>> >> >>> >> >>> >> >>> ____________________ >> >>> Racket Users list: >> >>> http://lists.racket-lang.org/users >> >> >> >> >> >> >> >> -- >> >> Jay McCarthy <[email protected]> >> >> Assistant Professor / Brigham Young University >> >> http://faculty.cs.byu.edu/~jay >> >> >> >> "The glory of God is Intelligence" - D&C 93 >> > >> >> >> >> -- >> Jay McCarthy <[email protected]> >> Assistant Professor / Brigham Young University >> http://faculty.cs.byu.edu/~jay >> >> "The glory of God is Intelligence" - D&C 93 >> >> ____________________ >> Racket Users list: >> http://lists.racket-lang.org/users > > -- Jay McCarthy <[email protected]> Assistant Professor / Brigham Young University http://faculty.cs.byu.edu/~jay "The glory of God is Intelligence" - D&C 93 ____________________ Racket Users list: http://lists.racket-lang.org/users

