Serge Abiteboul and Paris C. Kanellakis. Object identity as a query language primitive. SIGMOD Rec., 18(2):159--173, 1989.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....Pom Alg , to denote the restriction of Pom Alg without the operation . We next study the properties and dependencies of the different operations. We first consider the duplicate elimination operator. Note that this operator received much attention in the context of object oriented languages [AK89, BP91] It was shown in particular that it plays an important role to get complete languages. This operation is also very significant in the context of ordered types. The next propositions shows the limitations of Pom Alg ffl n . For that we introduce the notion of series parallel pomsets. ....

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. In Proc. ACM SIGMOD Int. Conf. on Management of Data, pages 159--173, 1989.

.... full spectrum of temporal constructs Employee s Name(String) s PaySlipNumber(Integer) Salary(Integer) vt Project ProjectCode(String) Manager vt TopManager AreaManager Department s InterestGroup OrganisationalUnit s d DEX Works for vt Manages Resp for s (1,n) act emp man (1,1) [1,5] prj (1,1) 1,n) prj org Fig. 1. A temporal ER diagram. cf. 26] for an extensive overview of temporally extended ER models) We refer to ER models because they are the most mature field in temporal conceptual modelling. As far as the notation for temporal constructs is concerned, we use a ....

....Depending on the application requirements, the temporal support must be decided by the designer. Furthermore, snapshot reducibility [39] of a schema says that snapshots of the database described by a temporal schema are the same as the database described by the same Orange Generate g Juice [1,1] [5,5] Fig. 2. Generation relationships. schema where all the temporal constructs are eliminated and the schema is interpreted atemporally. Temporal marks are usually added to capture the temporal behaviour of the di#erent components of a conceptual schema. In particular, entities, ....

[Article contains additional citation context not shown here]

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. Journal of the ACM, 45(5):798--842, 1998. A first version appeared in SIGMOD'89.

....among all ID attributes. An #####(#) annotation indicates a reference, i.e. it should contain a (set of) value(s) of some ID attribute(s) present in the document. Observe that the ID IDREF mechanism is similar to both the object identity based notion of references from object oriented databases [3] and to keys foreign keys from relational databases. On the one hand, like object identifiers, ID attributes uniquely identify elements within the whole document. On the other hand, as XML has a textual format, the reference semantics is achieved with implicit constraints that must hold on ....

.... ######### ###### ###### ######## ######### ###### ### ## ######### ####### ###### ######### ######### #### ######## ######### #### ### ## ######### ####### ##### ######### ######### ###### ######### Here the original ID semantics is appropriate to capture the notion of object identifiers [3] (the ### attributes) However, references through IDREF are rather weak: as IDREF attributes are untyped , we no longer know that the ## #### attribute of a ###### element should reference a department (a #### element) In addition, as in the previous example, keys are not precisely captured ....

S. Abiteboul and P. C. Kanellakis. Object identity as a query language primitive. In Proceedings of ACM SIGMOD Conference on Management of Data, pages 159--173, Portland, Oregon, June 1989.

.... of oids we do not make any assumptions about their nature and we invent fresh new base types for them (see Doid for Dept; we abused the notation a little by choosing for the dictionary the same name as the class) This representation actually corresponds to the usual semantics of OODB constructs [AK89] Sethstringi DProjs; string MgrNamegi Proj : SethStructfstring PName; string CustName; To complete the translation of the extended ODMG schema of figure 1.5 into our logical schema representation we need to represent, in addition to the schema names, the referential integrity (RIC) inverse ....

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. In Proceedings of ACM SIGMOD Conference on Management of Data, pages 159--173, Portland, Oregon, 1989.

.... do not make any assumptions about their nature and we invent fresh new base types for them (see Doid for Dept in figure 3; we abused the notation a little by choosing for the dictionary the same name as the class) This representation actually corresponds to the usual semantics of OODB constructs [1]. The syntax of queries and that of query plans are very close: for example, if d is an oid in depts the implicit dereferencing in d:DName corresponds to the dictionary lookup in Dept [ d ] DName. The relation Proj, stored as a table (a set of records) is also part of the physical schema, who ....

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. In Proceedings of ACM SIGMOD Conference on Management of Data, pages 159--173, Portland, Oregon, 1989.

....schemata. The proposal is based on a formal framework which consists of an object oriented data model extended with schema versions and its encoding in a description logics which supports complex reasoning tasks. The proposed model extends the snapshot object oriented model introduced in [1] and developed in [8] which is consistent with all the static features of UML OMT and ODMG models. Moreover, it provides full support for the taxonomy of primitive schema changes usually considered in the literature [4] Formal semantics for the schema (version) and for the supported schema ....

....the semantics of the underlying data model for the single schemata. 3.1 Syntax and Semantics The object oriented model we propose allows for the representation of multiple schema versions. It is an expressive version of the snapshot i.e. single schema object oriented model introduced by [1] and further extended and elaborated in its relationships with description logics by [8, 9] in this paper we borrow the notation from [8] The language embodies the features of the static parts of UML OMT and ODMG and, therefore, it does not take into account those aspects related to the ....

[Article contains additional citation context not shown here]

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. Journal of the ACM, 45(5):798--842, 1998. A first version appeared in SIGMOD'89.

.... paints properties) and they can be inherited (e.g. creates) while resources can be multiply classified (e.g. r1) In order to manipulate RDF (meta)data and schemas, we have to address the following issues: How can we capture the RDF typing primitives Clearly, standard object database models [3] fail to represent semistructured RDF (meta)data. On the other hand, most semistructured formalisms, such as OEM [14, 31, 30] are completely schemaless. Moreover, semistructured systems offering typing features (e.g. pattern instantiation) like YAT [17, 18] cannot exploit the semantics of data ....

.... we make the following contributions: ffl We introduce a graph data model bridging and reconciling RDF Model Syntax with Schema specifications [26, 10] The originality of our model relies on the distinction between classes and relationship types (e.g. properties) in the style of ODMG [3], as well as the introduction of a data graph instantiation mechanism inspired by GRAM [6, 7] ffl We propose a declarative language, called RQL, to query semistructured RDF (meta)data and related schemas. RQL adopts the syntax and functional approach of OQL [12] while it features generalized ....

[Article contains additional citation context not shown here]

S. Abiteboul and P. Kanellakis. Object Identity as a Query Language Primitive. In SIGMOD'89, pages 159--173, Portland Oregon, USA, June 1989.

....model [16, 17] In this paper, we deal with the introduction of temporal schema versioning in an object oriented database, also taking into account formal aspects. Within the object oriented framework, theoretical work has been done in the field of programming languages [18, 19] and databases [20, 21, 22], also including temporal ones [11] whereas thorough studies concerning schema versioning are still lacking. To this purpose, we will define (Object Oriented Data Model contextualized to chema ersions) that is a formal model for the management of temporal schema versioning in ....

....works and to the discussion of the approach in such a context. Conclusions can finally be found in Sec. 7. 2 Basic Definitions: Schema Versions and Database Instance is a temporal schema versioning model based on a generic object oriented data model, first introduced in [20], which is general enough to represent the static parts of UML and ODMG . In the following, starting from the definition proposed in [20, 24] we formally define the basic elements of . 2.1 Types, Subtyping, and Schema Versions In a database represents the structural ....

[Article contains additional citation context not shown here]

S. Abiteboul, P. Kanellakis, Object Identity as a Query Language Primitive, Journal of the ACM 45 (5) (1998) 798--842, a first version appeared in SIGMOD'89 Proceedings.

....However, the update itself may be defined by rules. Update programs have fixpoint semantics; the fixpoint can be computed by a bottomup evaluation according to a certain stratification. 1 Introduction databases has been discussed and investigated for a certain time now [Ban86, KNN89, Abi90, AK89, KL89, KLW90, DKM91] Various approaches to incorporate updates into deduction have been proposed. However, only a few of these take object orientation Fakultat fur Informatik, TU Braunschweig, W 3300 Braunschweig, Germany Permission to copy without fee all or part of this material is ....

....an employee in a managerial position (isa empl=pos mgr) who had not yet been updated ( E ) and initiates a modify of his salary method ( mod[E] sal (S ;S ) The second rule modifies the salary of all employees who are no managers. Assume in our to be updated object base a manager phil who makes 4000 and has no superior, and an employee bob who makes 4200 and phil being one of his superiors. Surely we expect that the update (as a whole) leaves phil in the class hpe with a salary of 4600 and bob fired (i.e. no more an employee) This is indeed the case (cf. figure 2) The first rule ....

[Article contains additional citation context not shown here]

Serge Abiteboul and Paris C. Kanellakis. Object identity as a query language primitive. In Clifford et al. [CLM89], pages 159 -- 173.

....of the other graphical primitives. Similarly, Figure 2 represents a fragment of the Louvre Web server [1] 3.2 Instances as Graphs We can now devote our attention to the instance level of ADM. Instances of our schemes can be defined in a way similar to those common for complex object data models [7, 21]. However, we prefer to show an alternative (but equivalent) view based on graphs; this will turn out to be useful in the next section, where we discuss the notion of navigation. The basic idea is that each page is GalleryPage URL cgi bin htima Artists Index ArtistsPage URL ....

....URL( Title ) specifies that the system has to generate an URL for each page over scheme BotticelliPaintingPage, and that the URL must be uniquely associated with the title value. 5 5This technique is somehow similar to the use of Skolem functors to invent new OID s in object oriented databases [7, 25, 22]. 15 In this way we can connect the query result page with the associated paintings pages. Here, we are implicitly supposing that a different URL can be generated for each painting, i.e. that the title uniquely identifies the painting. We now want to define more precisely the semantics of ....

[Article contains additional citation context not shown here]

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. In A CM SIGMOD International Conf. on Management of Data, pages 159-173, 1989.

....Another research direction is that of employing logical languages as query languages for objectoriented databases, leading to the definition of object oriented logical languages primarily used for querying object oriented databases. The most significant example of such a language is IQL [3]. The last direction we consider is that of extending Datalog to specify object oriented databases. In such extensions, the logical component is used to specify the schema of the database, and a distinction is made between base relations and derived relations. Examples of such systems are LOGRES ....

....of structured values. In COL besides types obtained by applying tuple and set constructors, union types are considered, since they support the manipulation of heterogeneous sets. The language does not consider neither object identity nor inheritance. IQL IQL (Idettity Query Latguage) [3] is an extension of COL with the concepts of object identity and inheritance. It is centered on the concept of object identifier (old) as a powerful primitive for database query languages. Informally, olds in IQL are typed pointers. Olds are used for three main purposes: i) representing ....

[Article contains additional citation context not shown here]

S. Abiteboul and P. Kanellakis. Object Identity as a Query Language Primitive. In Pvoc. of the ACM IGMOD Lnt'l Conf. on Management of Data, pages 159 173, 1989. 29

....and Atkinson et al. in the ObjectOriented Database System Manifesto [6] list a number of features which they argue are essential for OO database systems. Parallel to this, a number of formalizations of OO models have been proposed, such as COL [1, 2, 3] F logic [33] HILOG [13] ILOG [29] IQL [4, 5], formalizations of 02 [36, 35] OBJ and related languages [20, 21, 22, 23, 25, 38] OBLOG [15, 19, 17, 43] and CMSL [50, 46] Beeri [9] gives a survey of some issues. Even though these papers are top down, there is not yet any root concept from which they develop their formalization of ....

S. Abiteboul and P.C. Kanellakis. Object identity as a query language primitive. Technical Report 1022, INRIA, 1989.

....and Atkinson et al. in the ObjectOriented Database System Manifesto [6] list a number of features which they argue are essential for OO database systems. Parallel to this, a number of formalizations of OO models have been proposed, such as COL [1, 2, 3] F logic [33] HILOG [13] ILOG [29] IQL [4, 5], formalizations of 02 [36, 35] OBJ and related languages [20, 21, 22, 23, 25, 38] OBLOG [15, 19, 17, 43] and CMSL [50, 46] Beeri [9] gives a survey of some issues. Even though these papers are top down, there is not yet any root concept from which they develop their formalization of ....

....The trees of possibilities they explore have different roots and are rapidly evolving into a forest as diverse as the implementations of OO databases that exist now. One important concept that has arisen from the forest of formalization efforts is the concept of an object identifier [4, 17, 29, 31, 34, 45], itself stemming from the concept of a surrogate [14, 27] In this paper, I try to develop a theory of objects, using the concept of object identifier (oid) as root from which to develop and formalize other concepts. The language defined in this process is called CMSL (Conceptual Model ....

S. Abiteboul and P.C. Kanellakis. Object identity as a query language primitive. In ACM-SIGMOD International Conference on Management ofData, pages 159 173, Portland, June 1989. ACM. Sigmod Record, Vol. 18.

....By analogy, methods in deductive objectoriented databases should be defined using rules and encapsulated in class definitions. Such methods should be invoked through instances of the classes as well. However, most existing deductive object oriented database languages, including F logic [9] IQL [1], Datalog meth [2] ROL [12] Datalog [8] do not allow rule based methods to be encapsulated in the class definitions. The main difficulty is that the logical semantics is based on programs that are sets of rules. If rules are encapsulated into classes, then it is not clear how to define ....

S. Abiteboul and P. C. Kanellakis. Object Identity as a Query Language Primitive. In Proceedings of the ACM SIGMOD International Conference on Management of Data, pages 159--173, Portland, Oregon, 1989.

....are used for similar purposes. However, the notion of updating object state does not fit well in object oriented extensions of logic programming. In addition, also approaches developed in the database field, like e.g. Fre94,GLR90,LO91] do not consider state evolution. Many of the approaches [AK89,GLR90,CCCR 89] moreover, do not consider the behavioral component of objects, that is, methods. We think that this is an important issue because it overcomes the dichotomy between data and operations of the relational model. Few proposals moreover, deal with behavioral inheritance and ....

S. Abiteboul and P. Kanellakis. Object Identity as a Query Language Primitive. In Proc. of the ACM SIGMOD Int'l Conf. on Management of Data, pages 159--173, 1989. 6 Actually, state evolution can be accommodated in F-Logic through Transaction logic

.... descriptor (u; c 0 ; d) where c is a sub concept of c 0 (inclusion semantics of isa) Note also that except for the (thesaurus) term attribute defined on a tree structured domain, our description model can be expressed by any object oriented database model (see for example the model of [1]) i.e. can be supported by any object oriented system. In the following section, we shall show that by an appropriate coding of thesaurus terms, we do not need anymore an explicit tree structure (the tree structure being hidden in the value of the term) which renders the description model fully ....

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. In Proc. of ACM SIGMOD Int. Conference on Management of Data, Portland, Oregon, June 1989.

....ce dernier principe est leve dans certains cas tres precis tels que l utilisation du langage de requetes. 2.2 D efinition formelle d un mod ele orient e objet Cette section presente la definition formelle d un modele particulier. Cette presentation emprunte beaucoup aux mod eles IQL [AK89] et O 2 [LRV88, LR89a, LR89b] La plupart des modeles orientes objet possedent les caracteristiques decrites dans cette section. 2.2.1 Types et hi erarchies de classes Les definitions formelles d objet, de type et de hierarchies de classes sont etroitement liees. Un objet consiste en la donnee ....

S. Abiteboul and P. Kanellakis. Object Identity as a Query Language Primitive. In ACM SIGMOD International Conference, June 1989. to appear in JACM.

....o ered by object oriented databases. Deductive object oriented databases have been proposed as the natural next step to overcome these shortcomings. In the past decade, a number of deductive object oriented database languages have been proposed, such as O logic [18] revised O logic [13] IQL [1], LOGRES [7] Datalog meth [2] CORAL [22] F logic [12] Rock Roll [3] Datalog [10] and ROL [15, 16] However, most of them are only structurally object oriented. Important behaviorally object oriented features such as methods and encapsulation common in object oriented database ....

S. Abiteboul and P. Kanellakis. Object Identity as a Query Language Primitive. In Proceedings of SIGMOD 98, pages 159-173.

....schemes makes them more expressive. Usually, however, they are limited to disjoint unions [KV93] or labeled unions such as variant records [CACS94] Because of this limitation the reasoning about these types remains simple [Car88, Rem89] We argue that general union types such as used in [LR89, AK89, CL94] are a useful extension and even arise naturally in data models without union types. We can use them, for instance, to model optional integer elds by specifying their type as (null int) where null is a special basic type with only one value viz. null. Moreover, we could de ne several kinds of ....

....: P ete 00 ; age : 28; sex : male 00 ] This problem is solved in the IQL data model but leads to a more complex de nition of the data model. Since it does not play any role in the rest of this paper we do not present the IQL solution here but refer the interested reader to the IQL paper [AK89]. 3 Equality and Subtyping In this section we will give reasoning rules for deciding subtyping and equality. We assume a xed scheme S consisting of hC; type; isai that is sound. Under this scheme a type 1 is a subtype of type 2 , written as 1 2 , if it holds for all instances hO; val; ....

S. Abiteboul and P.C. Kanellakis. Object identity as a query language primitive. In ACM SIGMOD'89, pages 159-173, 1989.

....and completeness of theproposed resolution procedure were proven, along withanequivalentmodel theoretic semantics. The first attempt todevelopadeductivelanguage over an object oriented datamodel within a traditional database frameworkhas been pursued in a seminal paper byAbiteboul and Kanellakis [3], withtheproposal of theIQLlanguage. It involves a datamodel with a clear distinction between database schemeand instance, where complex structures are builtbymeans of tuple and set type constructors. The rule based language is a suitable 3 extension of Datalog for handling object identity.The ....

....is a set of clauses that specifies a transformation from an instance of the input schemetoaninstance of theoutputscheme. In order tokeep thesemantics as general as possible, we do not require disjointness between inputandoutputschemes (in contrast withother approaches, from Datalog [15, 28] to IQL[3]) Because of the presence of isa hierarchies, the usual separation between inputandoutputwould indeed be a limitation. Moreover, we do not require, as in other works [3, 10, 27] the presence of a most specific class for eachobjectofthedatabase, since this usually leads toanunreasonable increase ....

[Article contains additional citation context not shown here]

S. Abiteboul andP.Kanellakis. Object identity as a query language primitive. In ACM SIGMOD InternationalConf. on Management of Data, pages 159--173, 1989.

....transformations, but these approaches lack some typical OO features. COL (complex object language) AG91] extends Datalog by structured values and set constructors, without providing a class hierarchy or inheritance. Its semantics is given in terms of minimal models and stratification. IQL [AK92] provides oid s, set and tuple types corresponding to method applications, and types corresponding to classes where type inheritance corresponds to a subclass hierarchy. There is no value inheritance. IQL is evaluated bottom up using an inflationary fixpoint operator. LOGRES [CCCR 90] ....

S. Abiteboul and P. C. Kanellakis. Object Identity as a Query Language Primitive. In F. Bancilhon, C. Delobel, and P. Kanellakis, editors, Building an Object-Oriented Database System -- The Story of O2 , chapter 5, pages 98--127. Morgan Kaufmann, 1992.

....There are two ways around this. First, unranked trees can be encoded in a standard way as binary trees. Alternatively, the machinery and results developed for regular tree languages can be extended to the unranked case, as described in [20] an extension for unranked infinite trees is described in [5]) Either way, one can prove a surprising and satisfying connection between specialized DTDs and tree automata: they are precisely equivalent [20, 92] The equivalence of specialized DTDs and tree automata is a powerful tool for understanding XML schema languages. Properties of regular tree ....

S. Abiteboul and P. C. Kanellakis. Object identity as a query language primitive. JACM, 45(5):798--842, 1998. Extended abstract in SIGMOD'89.

....systems, etc. Though these two formalisms have been regarded as in opposition [26] indeed they can be integrated within a single approach, which selects compatible features and merges them into a unique setting. The feasibility of such an approach has been shown both at a theoretical [2, 18] and at an implementation level [8] This paper presents an overview of DEVICE, a DOOD that is based on the integration of deductive rules into an active OODB, by translating each deductive rule into special if then else production rules. Production rules are then emulated in the active OODB by ....

....this paper with a summary of the main points and a discussion of future work. 2. Related Work There are several approaches to the integration of deductive and object oriented databases [12] They mostly fall into three categories, a) object oriented (OO) extensions to Datalog, like e.g. [1, 2], b) OO extensions to Prolog, like e.g. 28, 21, 19] and c) OO extensions to first order logic programming [18] The most notable disadvantage of approaches in categories a) and c) is that they represent theoretical departures from logic, therefore their usability and practicality in terms of ....

S. Abiteboul and P.C. Kanellakis, Object identity as a query language primitive, Proc. Int. SIGMOD Conf. on the Management of Data, 1989.

....natural to replace every element of U U occurring in the output of a query applied to a database D by a fresh new element in U not in jDj. The class of abstract database transformations that can be obtained from computable generic queries in this way has a purely algebraic characterization [2, 64]. 5 Constraint databases Until now we have worked with relational databases over an unstructured universe U of atomic data elements. That the atomic data elements remain abstract and uninterpreted is one of the identifying features of classical database theory, and corresponds in practice to ....

S. Abiteboul and P.C. Kanellakis. Object identity as a query language primitive. Journal of the ACM, 45(5):798-842, 1998.

....IN, USA Abstract Various non deterministic aspects of object creation in database transformations are discussed, from a modeling as well as from an expressive power point of view. 1 Introduction In the past few years, a lot of attention has been paid to database transformations [AK89, AV88, AV90, AV91a, GPVG90, Hul86, HWWY91, HY91] Database transformations are binary accessibility relationships between database instances, and provide a simple, unifying model for the study of database query languages and important notions in database dynamics, such as updates, updatable ....

....in the input. Object creation differs from conventional insertion of a new value in that it contains an implicit form of non determinism, i.e. the particular choice of the new domain value to be added is irrelevant, as long as it is new. This weak form of non determinism is called determinacy [AK89] Object creation finds an obvious application in object oriented database systems, where, if a new object has to be added, the identity of this object must be distinct from those of all objects already present, but the particular choice of this new identifier is an implementation detail. ....

[Article contains additional citation context not shown here]

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. In J. Clifford, B. Lindsay, and D. Maier, editors, Proceedings of the 1989 ACM SIGMOD International Conference on the Management of Data, volume 18:2 of SIGMOD Record, pages 159--173. ACM Press, 1989.

....for the computation of domain preserving functions from relational structures to relations. But later, the focus shifted to the computation of graph functions, in the new context of object oriented databases. Indeed, various manipulations of object databases can be modeled as graph manipulations [1, 5]. Concretely, we will analyze the relationship between the GGM model and the two object database graph languages GOOD [11] and DMS [9] GOOD is a graph transformation language based on pattern matching, while DMS is an object oriented language based on the parallel invokation of methods. We prove ....

....be able to add new nodes to a graph. However, motivated by recent applications of databases in object oriented environments, the theory of database queries has recently been extended beyond domain preserving functions, allowing for the introduction of new domain elements in the result of a query [1, 16]. This extended theory, needed to account for the creation of new objects in an object oriented database, is fully applicable to the particular case of graphs. A powerful language for expressing object creating queries is the language provided by the GOOD model, a graph oriented model for object ....

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. In J. Clifford, B. Lindsay, and D. Maier, editors, Proceedings of the 1989 ACM SIGMOD International Conference on the Management of Data, volume 18:2 of SIGMOD Record, pages 159--173. ACM Press, 1989.

....[4] ensures that queries can be computed in a way independent of the encoding of the universe of possible domain values. More recently, object oriented database applications motivated researchers to relax domain preservation by allowing the appearance of new objects in the result of a query [2]. Extending the notion of query in this sense, however, necessitates a significant revision of the definition of Chandra and Harel. Indeed, queries which introduce new objects in their results are necessarily non deterministic due to the genericity criterion. Obviously, there is a need to control ....

....introduce new objects in their results are necessarily non deterministic due to the genericity criterion. Obviously, there is a need to control the degree of non determinism one wishes to allow. An important proposal to revise the definition of Chandra and Harel came from Abiteboul and Kanellakis [2]. They introduced the notion of determinate The following extended abstract presents research results of the Belgian Incentive Program Information Technology Computer Science of the future, initiated by the Belgian State Prime Minister s Service Science Policy Office. The scientific ....

[Article contains additional citation context not shown here]

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. In Proc. 1989 ACM SIGMOD Int. Conf. Management of Data, in SIGMOD Rec., 18(2):159--173, 1989.

....is applied. In order to be able to exploit this potential, we need at least be able to generate sets of new statement numbers. A practically convenient approach to accomplish this is to use an invention mechanism similar to the one used in update languages [2] and object oriented query languages [1, 11]. In our algebraic context, we define this mechanism as an additional algebra operator on relations as follows: 8 R = A a b c X 4 = N A sno 11 a 1 12 a 2 13 b 1 14 b 2 15 c 1 16 c 2 Figure 3: Illustration of identifier generating program from Example 2.7 Definition 2.6 Let be a ....

....just shown that its use for reification purposes can be circumvented. It should be pointed out, however, that value invention operations like the number operator also have other important uses, e.g. in obtaining computationally complete query languages and in object oriented query languages [1, 2, 11]. We will not consider these aspects in the present paper. 4.2 Query languages and complexity Two relational query languages will play a key role in our further analysis of the expressiveness and complexity of reflection: BA, the relational algebra extended with bounded looping, and WA, the ....

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. In Clifford et al. [8], pages 159--173.

....certain parts of the view schema to be hidden to the user of the view. This is desirable, e.g. for auxiliary methods used in the construction of the view. A good example is the method old risk in Example 3.1. Object creation: This is well understood from the theory of OO query languages (e.g. [AK89, VdBVGAG97]) one creates a new object as a function of a tuple of existing objects. For example, consider Family objects created as a function of (father; mother) pairs. We can easily model such an example by augmenting the root database, before actual view construction, with an intermediate layer holding ....

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. In J. Clifford, B. Lindsay, and D. Maier, editors, Proceedings of the 1989 ACM SIGMOD International Conference on the Management of Data, volume 18:2 of SIGMOD Record, pages 159--173. ACM Press, 1989. 12

....intermediate relations is an alternative to the intermediate relations with unbounded arity of QL. Motivated by applications in object oriented database systems [19] the need arose for queries where new objects do appear in the final result. This lead Abiteboul and Kanellakis to the study of IQL [4], a query language roughly comparable in expressive power to detTL but without the prohibition of new objects in the result. In an attempt to capture the queries expressible in IQL, they defined the intuitively appealing class of determinate queries. 2 A non deterministic query is called ....

....database. Hence, the isomorphism can only permute the new objects. Determinacy thus intends to isolate the very weak form of non determinism which is needed to accommodate new objects. It turned out that IQL is not complete for the determinate queries: an extra operation called copy elimination [4] has to be added in order to obtain determinate completeness. Nevertheless, most likely all determinate queries arising in practice can be already be expressed in IQL (without copy elimination. So, one could conclude that the class of determinate queries is perhaps not so natural after all. This ....

[Article contains additional citation context not shown here]

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. Technical Report 1022, INRIA-Rocquencourt, 1989. Also in SIGMOD Record, 18(2):159--173, 1989.

....and manipulation of set values, turned out to be important to support objectoriented applications of database systems [11] To deal with this new feature, a number of new data models and query languages were proposed. Particularly influential was the work of Abiteboul and Kanellakis [4]. They proposed the language IQL, which provides the necessary mechanisms for object creation and the representation of set values. In order to assess the expressiveness of IQL, Abiteboul and Kanellakis first had to extend the notions of genericity and completeness to database transformations ....

....Q 1 of Example 2.11. essentially deterministic, the programmer does not know in advance which cells will be allocated as they are chosen by the system. The desire to study database transformations involving object creation, but having a deterministic effect, have led Abiteboul and Kanellakis [4] to introduce the term determinate transformation for a database transformation in which any non determinism solely stems from the particular choice of new domain elements. Example 2.11 Let G be a binary relation name. An instance I of fGg can be interpreted as a directed graph with set of nodes ....

[Article contains additional citation context not shown here]

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. In J. Clifford, B. Lindsay, and D. Maier, editors, Proceedings of the 1989 ACM SIGMOD International Conference on the Management of Data, volume 18:2 of SIGMOD Record, pages 159--173. ACM Press, 1989. Also INRIA Rapport de Recherche 1022, 1989.

....given logic formula, thus acting as a filter. Selection operations are of fundamental importance in all high level database management tasks. Inspired by this, an important issue in the research on next generation and objectoriented databases is the development of formal, logic based data models [AK89, Bee90, KLW95, LR89, SS90] which allow the design of high level query languages (a good example is [BCD89] which is based on [LR89] Recent efforts in this area typically provide dynamic features (like object identity, inheritance, and methods) on top of a core complex object data model. Complex ....

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. In Clifford et al. [CLM89], pages 159--173.

No context found.

S. Abiteboul and P . C. Kanellakis. Object identity as a query language primitive. In Proc.ofACM SIGMOD, 1989.

No context found.

S. Abiteboul and P . C. Kanellakis. Object identity as a query language primitive. In Proc. of ACM SIGMOD, 1989.

....as the natural next step to overcome these shortcomings (see for instance the proceedings of the DOOD conferences) We consider an extension of datalog with classes, methods, inheritance and a view mechanism and study method resolution for this language. The contribution is in the spirit of [AK89, KL89] a new (we believe important) step towards flexible and formal languages for databases. A major issue in object oriented languages is that of INRIA, 78153 Le Chesnay, France, Serge.Abiteboul inria.fr, Emmanuel.Waller inria.fr. Work partially supported by Esprit BRA Project Fide2. ....

....body. Second we restrict the way classes are populated. In the following, we assume that for each oid o in Delta, there is exactly one ground fact c(o) in Delta. The class c is called the base class of an object o, if fact c(o) is in Delta. This is in the spirit of the oid assignment of IQL [AK89] The following notions to deal with the class hierarchy will be used. Let the transitive reflexive closure of . The set of all super and subclasses of a given class name c is defined as c j c c g, and c j c cg, respectively. Let D C be a set of class names. The subset of ....

S. Abiteboul and P. C. Kanellakis, Object Identity as a Query Language Primitive", Proc. SIGMOD, 1989, to appear in J. ACM.

....specification task and also helps in preventing inconsistencies in specifications. It should be noted that the language we use to define correspondence rules is rather limited. Similar correspondences could be easily derived using more powerful languages previously proposed (e.g. LDL [8] or IQL [7]) But in these languages it would be much more difficult (sometimes impossible) to derive translation rules from given correspondence rules. Nevertheless, our language is expressive enough to describe many desired correspondences translations. As will be seen, correspondence rules have a very ....

....(This representation is clearly redundant and an alternative representation that would avoid repeating the type can easily be specified. An object oriented database can be viewed as a cyclic graph. However, using object identifier one may easily represents a cyclic graph as a tree, see, e.g. [7]. We consider here the ODMG model and a possible representation for it. An object is represented by the root vertex of a tree representing the value of the object as follows: ffl For an atomic value, we use a vertex labeled by the atomic type and whose unique child is labeled by the appropriate ....

[Article contains additional citation context not shown here]

S. Abiteboul and P. C. Kanellakis. Object identity as a query language primitive. In Proc. of the ACM SIGMOD Conf. on Management of Data, pages 159--173, 1989.

No context found.

Serge Abiteboul and Paris C. Kanellakis. Object identity as a query language primitive. SIGMOD Rec., 18(2):159--173, 1989.

No context found.

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. Journal of the ACM, 45(5):798--842, 1998. A first version appeared in SIGMOD'89.

No context found.

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. Journal of the ACM, 45(5):798--842, 1998. A first version appeared in SIGMOD'89.

No context found.

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. Journal of the ACM, 45(5):798--842, 1998. A first version appeared in SIGMOD'89.

No context found.

S. Abiteboul, P. Kanellakis. Object identity as a query language primitive. J. ACM 45 (1998), 798--842.

No context found.

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. Journal of the ACM, 45(5):798--842, 1998. A first version appeared in SIGMOD'89.

No context found.

Serge Abiteboul and Paris Kanellakis. Object identity as a query language primitive. In Proc. of the ACM SIGMOD Int. Conf. on Management of Data, pages 159--173, 1989.

No context found.

Serge Abiteboul and Paris Kanellakis. Object identity as a query language primitive. In Proc. of the ACM SIGMOD Int. Conf. on Management of Data, pages 159-173, 1989.

No context found.

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. Journal of the ACM 45(5), pages 798--842, 1998.

No context found.

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. Journal of the ACM 45(5), pages 798--842, 1998.

No context found.

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. Journal of the ACM, 45(5):798--842, 1999.

No context found.

S. Abiteboul and P. Kanellakis. Object identity as a query language primitive. Journal of the ACM, 45(5):798--842, 1998. A first version appeared in SIGMOD'89.

No context found.

: S. Abiteboul, P. Kanellakis, Object identity as a query language primitive, SIGMOD Record (ACM Special Interest Group on Management of Data), n 18(2), pp. 159-173, Juin 1989.

No context found.

Abiteboul. S, Kanellakis. P, Object Identity as a Query Language Primitive, Proceedings of the ACM SIGMOD International Conference on Management of Data, pp. 159-173, 1989.

First 50 documents  Next 50

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