ABSTRACT
A number of optimizations have been proposed for Datalog programs involving a single intensional predicate (“single-IDB programs”). Examples include the detection of commutativity and separability ([Naug88],[RSUV89], [Ioan89a]) in linear logic programs, and the detection of ZYT-linearizability ([ZYT88], [RSUV89], [Sara89], [Sara90]) in nonlinear programs. We show that the natural generalizations of the commutativity and ZYT-linearizability problems (respectively, the sequencability and base-case linearizability problems) are undecidable. Our constructions involve the simulation of context-free grammars using single-IDB programs that have a bounded number of initialisation rules. The constructions may be used to show that containment (or equivalence) is undecidable for such programs, even if the programs are linear, or if each program contains a single recursive rule. These results tighten those of [Shmu87] and [Abit89].
- Abit89.Ablteboul, S {1989} "Boundedness ls undecldable for Datalog programs with a single recurslve rule," Informa~wn Processes9 Letters 3~, pp 281-287 Google ScholarDigital Library
- BMSU86.Bancflhon, F, D E Miner, Y Saglv and 3 D Ullman {1986} "Magic sets and other strange ways to implement logic programs," P~oc F~fth A CM Symposzum on Pmnczples of Database Systems, pp 1-15 Google ScholarDigital Library
- BR87.Been, C and R Ramaknshnan {1987} "On the power of magic," Proc S~zth A CM Symposium on Principles of Database Systems, pp 269-283 Google ScholarDigital Library
- CM77.Chandra, A K and P M Merhn {1977} "Ophmal xmplementatlon of conjunchve queries in relahonal databases," Proc N~n~h Annual A CM Symposium on $he Theory of Compu$- zng, pp 77-90 Google ScholarDigital Library
- HU79.Hopcroft, :I E and J D Ullman {1979} In~roduc$zon ~o automata ~heory, languages and compu~a$zon, Addxson-Wesley Google ScholarDigital Library
- Ioan89a.Ioanmd,s, Y E {1989} "Commutahv,ty and its role m the processing of hnear recursmn," Tech Report 804, Umverslty of Wisconsin- MadisonGoogle Scholar
- Ioan89b.Ioanmdls, Y E {1989} "Towards an algebrazc theory of recursmn," Tech Report 801, Umverslty of Wisconsin-MadisonGoogle Scholar
- Naug88.Saughton, J F {1988} "Compiling separable recursmns," A CM SIGMOD Intl Conf on the Managernen~ of Data, pp 312-319 Google ScholarDigital Library
- RSUV89.Ramaknshnan, R, Y Saglv, J D Ullman and M Y Vardl {1989} "Proof-tree transformahon theorems and their apphcatmns," Proc E~ghth A CM Symposium on Pmnczples of Database Systems, pp 172-181 Google ScholarDigital Library
- SY81.Sagav, Y and M Yannakahs {1981} "Eqmvalence among relational expressions with the union and difference operators," J A CM $7, pp 633-655 Google ScholarDigital Library
- Sagi87.Saglv, Y {1987} "Optimizing Datalog programs," P.oc S~zth A CM Symposium on Prznc~ples of Database Systems, pp 349-362 Google ScholarDigital Library
- Sara89.Sarmya, Y {1989} "Lmeanzmg nonhnear recursmns m polynormal hme," Proc E~gh~h A CM Symposium on Pmnc~ples of Database Systems, pp 182-189 Google ScholarDigital Library
- Sara90.Sarazya, V {1990} "Polynommlhme program transformatmns m deduchve databases," to appear m Proc N~nth A CM Symposium on Pmnczples of DaSabase Sys- Zeros Google ScholarDigital Library
- Shmu87.Shmueh, O {1987} "Dec~dabxhty and expressxveness aspects of logxc querms," Proc Szz~h A CM Symposium on Pmnczples of Database SysSems, pp 237-249 Google ScholarDigital Library
- Ullm89.Unman, ~ D {1989} Prmczples o/Da~.abase and Knowledge.base Systems, Vol II, Computer Scmnce Press, Rockwlle, MdGoogle Scholar
- ZYT88.Zhang, W, C T Yu and D Troy, "A necessary and sufficmnt cond~tmn to hneax~ze doubly recurs~ve programs m logic databases," unpubhshed manuscript, Dept of EECS, Umvers~ty of Ilhnms at ChicagoGoogle Scholar
Index Terms
- Hard problems for simple logic programs
Recommendations
Hard problems for simple logic programs
A number of optimizations have been proposed for Datalog programs involving a single intensional predicate (“single-IDB programs”). Examples include the detection of commutativity and separability ([Naug88],[RSUV89], [Ioan89a]) in linear logic programs, ...
Constructing the Least Models for Positive Modal Logic Programs
We give algorithms to construct the least L-model for a given positive modal logic program P, where L can be one of the modal logics KD, T, KDB, B, KD4, S4, KD5, KD45, and S5. If L ∈ {KD5,KD45,S5}, or L ∈ {KD,T,KDB,B} and the modal depth of P is ...
Comments