P. Atzeni and R. Torlone. Updating intensional predicates in datalog. Data and Knowledge Engineering, 7(8):1--17, 1992.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....the views are used. The user interacts with a view by issuing queries and update requests. The view definition mapping is sufficient to translate queries on views into queries on the underlying database. View updates, however, present a difficulty, referred to as non determinism in the literature [4, 3, 2, 11] which can be stated as follows: Given a view v and an update U on v, there exists in general more than one way to translate U into updates on the relations explicitly stored in the database. The objective of this paper is to study the view update translations in the context of the relational ....

P. Atzeni and R. Torlone. Updating Intensional Predicates in Datalog. Data and Knowledge Engineering, 8, 1992.

....deleting) a fact f over an intensional predicate, the set of extensional facts stored in the database must be modified so that f can be deduced (resp. cannot be deduced) from the updated database. This way of performing updates, based on abduction, has been considered in many approaches (see [3], 4] 10] Unfortunately, in many cases, there are several ways to modify the set of stored facts, and this is known as the non determinism of intensional updates (see [3] 12] In these cases, abduction is used in order to compute all the possibilities and it is up to the user (or the ....

....from the updated database. This way of performing updates, based on abduction, has been considered in many approaches (see [3] 4] 10] Unfortunately, in many cases, there are several ways to modify the set of stored facts, and this is known as the non determinism of intensional updates (see [3], 12] In these cases, abduction is used in order to compute all the possibilities and it is up to the user (or the system) to choose among them; when no choice is done the update is not performed. This shows that, in a sense, in these approaches, old data have priority over new data. For ....

Atzeni P., Torlone R., 1992. Updating Intensional Predicates in Datalog, Data & Knowledge Engineering, 8.

....In [35, 29] it is maintained that both database semantics and application semantics are required; some can be obtained at view de nition time, and the remaining must be obtained at view update request time. Some work has been done on the view update problem in databases which allow null values [6, 7, 28, 2]. Langerak [28] studied the view update problem for a relational data model in which the base relation may contain (indexed) nulls. Langerak considered an abstract class of views which are total projections of the representative instances (the representative instances are de ned to be the correct ....

....the basis for any insertion or deletion operation (modi cations are not considered as a separate class of updates) Di erent from Langerak [28] Vossen and Brosda do not discreminate database nulls from nondistinguished variables. The view update problems in deductive databases are considered in [3, 2, 8, 14, 15, 32, 21, 33]. 8 Summary In this paper, we study a new approach to view update translation. The basic idea is to transform the problem into constraint satisfaction problems. We provide sound and complete transformation for updates of PSUJ views. This is, to our knowledge, the rst reported attempt in this ....

P. Atzeni and R. Torlone. Updating intensional predicates in datalog. Data & Knowledge Engineering, 8(1):1{ 17, April 1992.

....updates of the underlying base (stored) facts. In general, several translations 3 may exist, and the user must select one of them. This problem has attracted much research during the last years in relational [BS81, CP84, Mas84, Kel85, Kel86, Dat86, Lan90, LS91] and in deductive databases [Tom88, Bry90, Dec90, KM90, GL90, GL91, TA91, TO92, AT92]. It has also been studied for logical databases [FUV83, FKUV86] In principle, it may happen that some translations corresponding to a given view update request do not satisfy the integrity constraints. For this reason, view updating is usually followed by an integrity checking process. The ....

....been proposed to solve this problem. The first one suggests treating views as abstract data types [FC85, MW88] so that the definition of a view includes all permissible view updates together with their translation. The second approach is to define a general translation procedure (a translator) [Tom88, Bry90, Dec90, KM90, GL90, GL91, TA91, TO92, AT92]. Inputs to the translator are a view definition, a view update request and the current knowledge base; and the output is a knowledge base update that satisfies the request. The method that we propose in this paper follows the translator approach. Usually, there exist several possible ways of ....

[Article contains additional citation context not shown here]

Atzeni, P.; Torlone, R. "Updating Intensional Predicates in Datalog", Data & Knowledge Engineering (8), 1992, pp. 1-17.

....de mise a jour est d eterministe, 2. elle se situe dans le cadre du mod ele relationnel, et 3. elle concerne les mises a jour sur les relations de base et sur les vues. Le probl eme de mises a jour des vues a fait r ecemment l objet de nombreux travaux, dans le contexte du mod ele relationnel [BS81, AT92a, LPS92, LS94]. Le probl eme, dans les approches traditionnelles ( BS81, AT92a, AT92b] est que, pour effectuer une mise a jour sur une vue, il faut traduire celle ci en termes de mises a jour sur la base de donn ees extensionnelle, et il est bien connu (voir notamment [AT92a] que cette traduction n est en ....

....relationnel, et 3. elle concerne les mises a jour sur les relations de base et sur les vues. Le probl eme de mises a jour des vues a fait r ecemment l objet de nombreux travaux, dans le contexte du mod ele relationnel [BS81, AT92a, LPS92, LS94] Le probl eme, dans les approches traditionnelles ([BS81, AT92a, AT92b]) est que, pour effectuer une mise a jour sur une vue, il faut traduire celle ci en termes de mises a jour sur la base de donn ees extensionnelle, et il est bien connu (voir notamment [AT92a] que cette traduction n est en g en eral pas unique. Dans le cadre des interfaces de type relation ....

[Article contains additional citation context not shown here]

P. Atzeni and R. Torlone. Updating Intensional Predicates in Datalog. Data and Knowledge Engineering, 8, 1992.

....each time the view is used. The user interacts with a view by issuing queries and update requests. The view definition mapping is sufficient to translate view queries into database queries. View updates, however, present a difficult problem, referred to as non determinism in the literature [BS81, AT92b, AT92a, LS94]. This problem can be stated as follows: Given a view v and an update U on v, there exists in general more than one way to translate U into updates on the relations explicitly stored in the database. The problem of view update naturally arises in data warehousing environments. A data warehouse is ....

P. Atzeni and R. Torlone. Updating Intensional Predicates in Datalog. Data and Knowledge Engineering, 8, 1992.

....research in the past [Mai83, Ull89] When thinking of deductive databases, in particular, people tend to identify the notion of inference with the problem of answering queries. However, there is a considerable amount of research concentrating on inference processes activated in response to updates [Bry90, CeW90, DaW89, DBM88, Dec86, NiY83, RSS90, SSRN91, ToA91, VBK91, AtT92]. Most of this work is devoted to a particular application of update driven inference, namely integrity checking over rule defined data. It is well known that there are other important problems similarly requiring computation of the consequences that an update on explicitly stored (base) data has ....

P. Atzeni, R. Torlone, "Updating Intensional Predicates in Datalog," Data & Knowledge Engineering, 8, 1992.

....how a request to update a view fact should be appropriately translated into updates of the underlying base facts. In general, several translations that satisfy a view update may exist. Much research has been devoted to this problem within the field of deductive databases; see for example [Dec90, KM90, GL91, AT92]. On the other hand, integrity constraints maintenance [ML91, W t93] consists of, given a consistent database state and a set of updates that violates some integrity constraint, determine a 12 set of additional base fact updates to be appended to the original update set such that the ....

Atzeni, P.; Torlone, R. "Updating Intensional Predicates in Datalog", Data & Knowledge Engineering (8), 1992, pp. 1-17.

....to recursive definitions. In this paper, we report on a new pragmatic approach to the definition and the implementation of updates to Datalog databases [8] which derives from a semantics of updates to relational databases in the weak instance model [4] and extends some previous results of ours [3]. An important difference between our proposal and the existing ones is that we take into account integrity constraints (specifically, functional dependencies) defined on extensional relations, both for consistency checking and for resolving possible ambiguities. Moreover, in our work we first ....

.... and we say that a Datalog database D over S is consistent if edb(D) satisfies the dependencies in fd (and so lhm(D) 3 Semantics of Update Operations In this section we present a declarative semantics of update operations in the framework of deductive databases which extends the one given in [3] taking into account the presence of integrity constraints. 3.1 Insertions Let S = ipred; epred; R) be a Datalog database scheme with associated a set of fd s fd, D be a consistent database over S, and f 2 hb be a fact such that D 6j= f . The semantics of the insertion of f in D can be defined ....

P. Atzeni and R. Torlone. Updating intensional predicates in datalog. Data and Knowledge Engineering, 7(8):1--17, 1992.

....to recursive definitions. In this paper, we report on a new pragmatic approach to the definition and the implementation of updates to Datalog databases [8] which derives from a semantics of updates to relational databases in the weak instance model [4] and extends some previous results of ours [3]. An important difference between our proposal and the existing ones is that we take into account integrity constraints (specifically, functional dependencies) defined on extensional relations, both for consistency checking and for resolving possible ambiguities. Moreover, in our work we first ....

.... and we say that a Datalog database D over S is consistent if edb(D) satisfies the dependencies in fd (and so lhm(D) 3 Semantics of Update Operations In this section we present a declarative semantics of update operations in the framework of deductive databases which extends the one given in [3] taking into account the presence of integrity constraints. 3.1 Insertions Let S = ipred; epred; R) be a Datalog database scheme with associated a set of fd s fd, D be a consistent database over S, and f 2 hb be a fact such that D 6j= f . The semantics of the insertion of f in D can be defined ....

P. Atzeni and R. Torlone. Updating intensional predicates in datalog. Data and Knowledge Engineering, 7(8):1--17, 1992.

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