Bancilhon, F., " note on the performance of Rule Based Systems" MCC Technical Report, 1985.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... where the set of nodes Me and the set of arcs E are defined as follows: a) NA,a ] where a denotes the given constant in the q th argument of the query goal, is in M and it is called the source node, moreover b) if IN h,b] is in MQ and c is in IN S(ND( b ) then [S (Nh ) c ] is in MQ and ([Nh ,b ], S (Nh ) c ] is in E. 18 The magic graph is a p partite graph, where p is the size of the active binding cycle, and it can be con structed in linear time. For each N h in the active binding cycle, let M denote all the nodes in M o that have Nh as first com ponent. M will be called a magic ....

....of the fit t ]eveb. However, since we already know that the q th column of the result only contai the value a, we can use a more efficient method which retur result the projection of on all columns but the q th. informal description of this method (called the counting method) can be found in [B]. Let us now present an algorithm for this method. In the algorithm shown below, by x[ x ] we denote the projection on all colum but thee named Counting Algorithm 1) Compute the counting sets M . M and 3) for vt,t 1, 2do begin 4) st g = 6) Compute R ] x,Xl g ,c , s( w ....

Bancilhon, F., " note on the performance of Rule Based Systems" MCC Technical Report, 1985.

....optimization of the bottom up approach for such programs. The bottom up approach consists of first applying optimizing program transformations and then evaluating the fixpoint of the rewritten program bottom up. It is generally proposed that the fixpoint be computed using the Seminaive algorithm [4, 7, 3], which avoids repeating inferences. However, this method requires that the new facts computed in each iteration be compared with previously generated facts to eliminate duplicates, and this is a costly operation, especially in the presence of non ground tuples. Eliminating duplicates now ....

....integrity constraints. We allow EDB facts to contain (universally quantified) variables. The treatment of bottom up evaluation of logic programs (or Datalog programs) in the literature has considered the evaluation as a computation of relations, and hence has used sets to describe the process [29, 4, 7, 3, 2]. Since our interest is in the occurrence of duplicate atoms (tuples) during the computation, we first extend portions of the existing theory to use multisets of atoms, rather than sets. We also extend most treatments by generating non ground atoms in a manner similar to [10] This involves a ....

F. Bancilhon, A Note on the Performance of Rule Based Systems. In MCC Technical Report DB-022-85, 1985.

....variable, since each of these could potentially restrict the computation of the answers. As we proceed, we have a collection of adorned predicates, and as ################ Seminaive fixpoint computation is a refinement that avoids repeating inferences in different iterations. See, e.g. [Ba85]. Note that non linear rules are also treated. 9 each one is processed, we will mark it, so that it will not be processed again. If p a is an unmarked adorned predicate, then for each rule that has p in its head, we generate an adorned version for the rule and add it to P ad ; ....

....the previous section, we showed that the Magic Templates algorithm transformed the given program into an equivalent program with respect to the query. The fixpoint of the transformed program is computed using a bottom up iteration, possibly with some refinements as in Seminaive evaluation (e.g. [Ba85]) In this section, we consider some properties of the fixpoint evaluation of the transformed program. 6.1. Optimality of the Magic Implemention of Sips Our main result concerns the optimality of the Magic Templates strategy, in the sense that it implements a given set of sips by computing the ....

[Article contains additional citation context not shown here]

F. Bancilhon, "A Note on the Performance of Rule Based Systems," MCC Technical Report DB-022-85, 1985.

....here refers to instantiation of all subgoals with established facts so that all constraints in the rule body are satisfied. An evaluation has the semi naive property if no rule firings in different iterations are duplicated. The basic semi naive evaluation was rediscovered by several researchers [B, BalR1, Bayer] and was generalized and improved a few times [BalR2, KNSS, RSS] Essential to these variations of the semi naive evaluation are rewriting each rule into a number of differential or seminaive versions and firing them through use of relational join operation. Recently, the following observation ....

F. Bancilhon, "A note on the performance of rule based systems," MCC Technical Report DB-022-85

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC