Abiteboul, S., Cluet, S., and Milo, T. (1997). Correspondence and Translation for Heterogeneous Data. In Intl. Conf. on Database Theory. Delphi, Greece.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....the corresponding skeleton plan to execute the query. The results we develop here complement their work by identifying the set of mediator query forms for which skeleton plans may be precomputed. Related research has explored heterogeneous database schema integration and data translation [1, 8, 49], as well as capability based query optimization [20, 57, 80] The techniques we developed to compute mediator capabilities complement those efforts. In data integration systems, it is important to describe the capabilities of mediators so that they can be used as easily (by end users as well as ....

S. Abiteboul, S. Cluet, T. Milo. Correspondence and Translation for Heterogeneous Data. In Proc. ICDT Conference, 1997.

....we were guided mainly by our intuition rather than by formal considerations. It should be noted that there is a long history of work on data translation that focuses on taking high level, data independent translation rules and generating efficient, executable translation programs [SHL75,SHT 77,ACM97] Yet, we could not find a formal justification for the intuitive choices we made in creating the target instance. In seeking to formalize this intuition and justify the choices made in Clio, we were led to explore foundational and algorithmic issues related to the semantics of data exchange and ....

S. Abiteboul, S. Cluet, and T.. Milo. Correspondence and Translation for Heterogeneous Data. In Proceedings of the International Conference on Database Theory (ICDT), pages 351--363, 1997.

....we were guided mainly by our intuition rather than by formal considerations. It should be noted that there is a long history of work on data translation that focuses on taking high level, data independent translation rules and generating efficient, executable translation programs [SHL75,SHT 77,ACM97] Yet, we could not find a formal justification for the intuitive choices we made in creating the target instance. In seeking to formalize this intuition and justify the choices made in Clio, we were led to explore foundational and algorithmic issues related to the semantics of data exchange and ....

S. Abiteboul, S. Cluet, and T.. Milo. Correspondence and Translation for Heterogeneous Data. In Proceedings of the International Conference on Database Theory (ICDT), pages 351--363, 1997.

....related idea, proposed by Wadler [32] extended algebraic pattern matching to abstract data types using programmer supplied in and out operators. This is essentially the special case of our lenses in which the get and put functions must always form an isomorphism. Abiteboul, Cluet, and Milo [2] defined a declarative language for describing correspondences between parts of trees in a data forest. In turn, these correspondence rules can be used to translate one tree format into another through non deterministic Prolog like computation; however, this process requires an isomorphism between ....

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proceedings of 6th Int. Conf. on Database Theory (ICDT), 1997.

....related idea, proposed by Wadler [30] extended algebraic pattern matching to abstract data types using programmer supplied in and out operators. This is essentially the special case of our lenses in which the get and put functions must always form an isomorphism. Abiteboul, Cluet, and Milo [2] defined a declarative language for describing correspondences between parts of trees in a data forest. In turn, these correspondence rules can be used to translate one tree format into another through nondeterministic Prolog like computation; however, this process requires an isomorphism between ....

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proceedings of 6th Int. Conf. on Database Theory (ICDT), 1997.

....instance, we were guided mainly by our own intuition rather than by formal considerations. It should be noted that there is a long history of work on data translation that focuses on taking high level, data independent translation rules and generating efficient, executable translation programs [1,20,21]. Yet, we could not find a formal justification for the intuitive choices we made in creating 210 R. Fagin et al. the target instance. In seeking to formalize this intuition and justify the choices made in Clio, we were led to explore foundational and algorithmic issues related to the semantics ....

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and Translation for Heterogeneous Data. In ICDT, pages 351--363, 1997.

....such as XQuery [8] implement only approximate type checking) Worse still, any approach based on type checking does not provide any guidance on how to define an XML view that does typecheck. There has also been a line of work on automated inference of mappings from schema information (e.g. [2, 4, 17]) This approach works well when the source and target schemas involved in translations are similar to each other. If the schemas are dramatically different, or if the view mapping depends on the application rather than merely upon the schemas, this process cannot be fully automated. In this ....

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In ICDT, 1997.

....guided mainly by our own intuition rather than by formal considerations. It should be noted that there is a long history of work on data translation that focuses on taking high level, data independent translation rules and generating efficient, executable translation programs [SHL75, SHT 77, ACM97] Yet, we could not find a formal justification for the intuitive choices we made in creating the target instance. In seeking to formalize this intuition and justify the choices made in Clio, we were led to explore foundational and algorithmic issues related to the semantics of data exchange and ....

S. Abiteboul, S. Cluet, and T.. Milo. Correspondence and Translation for Heterogeneous Data. In Proceedings of the International Conference on Database Theory (ICDT), pages 351--363, 1997.

....instance, we were guided mainly by our own intuition rather than by formal considerations. It should be noted that there is a long history of work on data translation that focuses on taking high level, data independent translation rules and generating efficient, executable translation programs [1, 20, 21]. Yet, we could not find a formal justification for the intuitive choices we made in creating the target instance. In seeking to formalize this intuition and justify the choices made in Clio, we were led to explore foundational and algorithmic issues related to the semantics of data exchange and ....

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and Translation for Heterogeneous Data. In ICDT, pages 351--363, 1997.

....such as XQuery [8] implement only approximate type checking) Worse still, any approach based on type checking does not provide any guidance on how to define an XML view that does typecheck. There has also been a line of work on automated inference of mappings from schema information (e.g. [2, 4, 17]) This approach works well when the source and target schemas involved in translations are similar to each other. If the schemas are dramatically different, or if the view mapping depends on the application rather than merely upon the schemas, this process cannot be fully automated. In this ....

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In ICDT, 1997.

....context, Vassalos and Papakonstantinou introduce in [VP97] a higher order language termed RQDL suitable for describing data integration inside a mediated architecture. They describe a translation process of that language into a first order logic language noted p Datalog, with function symbols. In [ACM96], Abiteboul, Cluet and Milo, introduce a system suitable for integrating object oriented databases. They introduce a logic language capable of describing some tree transformations for object oriented schema integration. In addition to these efforts, a vast amount of schema integration systems and ....

....[VRG98] among others or [FLM98] for a general overview. The UniSQL language is described in [KGK95] and introduces operators to process dynamic schema updates. The SIMS system proposed by Arens et al. see [AK93] AKS96] is a query rewriting system for rederating databases. 23 References [ACM96] S. Abiteboul, S. Cluet, T. Milo: Correspondence and Translation for Heterogeneous Data . AK93] Arens, Knoblock. SIMS: Retrieving and Integrating Information From Multiple Sources . Proceedings of the ACM SIGMOD Conference 1993 [AKS96] Y. Arens, C. Knoblock, W. Shen: Query Reformulation for ....

[Article contains additional citation context not shown here]

S. Abiteboul, S. Cluet, T. Milo: "Correspondence and Translation for Heterogeneous Data".

....in [VP97] a higher order language termed RQDL suitable for describing data integration inside a mediated architecture. They describe a translation process of that language into a first order logic language noted p Datalog, in which skolem constants play the same role as in our construction. In [ACM96], Abiteboul, Cluet and Milo, introduce a system suitable for integrating object oriented databases. They introduce a logic language capable of describing some tree transformations for object oriented schema integration. In addition to these efforts, a vast amount of schema integration systems and ....

S. Abiteboul, S. Cluet, T. Milo: "Correspondence and Translation for Heterogeneous Data".

....the document interpreter is to gather and combine information from the data sources, as instructed in the document prescription, and to map it according to the document DTD. The integration of heterogeneous information is done using a tree like model which is very similar to those presented in [5] or [4] As in these approaches we are not using a model that encompasses all the source models. Instead we use a generic and minimalist approach. The data structure we use consists of ordered As this page is generated internally, copyright is not an issue here. We realize however that this is an ....

....is not permitted: the start tag and the corresponding end tag must both be present. XML makes the difference between a valid document, which is a document conform to its DTD, and a well formed document which only requires tags to Of course this mapping is not unique; for an alternative see [5] be properly embedded. This is extremely important since it allows to process a document without its DTD, and to transmit and exchange fragments of documents without parsing the full document. XML offers a set of constructs for describing links between objects: in addition to the target ....

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proceedings of ICDT97, 1997.

....works and provides some comparisons with existing languages. 2. Information Objects In order to integrate information objects from heterogeneous sources, we define an object model which provides a common view of all information objects. This model is very similar to the models presented in [ACM97] or [Abi97] As in these approaches we are not using a model that encompasses all the data sources models: instead we use a generic and minimalist approach. In this section we present our object model and show how some information objects sources can be represented using this model. The object ....

....or within SGML editors for supporting cut and paste between incompatible elements [AQR94] Other languages for restructuring information, like Lorel [AQM96] focus on the integration of heterogeneous data structures or translation from one world to another, including from to a document format. [ACM97] defines a language for correspondence between schemas and data in the style of rule based systems. These languages, as well as DSSSL, could be used for implementing library coercion functions to map complex or frequently used structures, but are beyond the reach of document writers because of ....

S. Abiteboul, S. Cluet, and T. Milo. "Correspondence and translation for heterogeneous data", in Proc. ICDT 97.

....assumes that actual prices are listed in purchase order, while cXML assumes that the prices are taken from product catalog. Theoretically speaking, the problem of direct translation of the trees encoded with XML is quite complicated and requires a number of significant restrictions to be solvable [Abiteboul et al., 2001]. In addition to high document complexity, a marketplace faces the problem of complex (and exponential in their number) mappings between the objects. In consequence, the rules, which try to carry out the complete transformation process in one shot, have proven to be very complex. This causes ....

Abiteboul, S., Cluet, S., and Milo, T., "Correspondence and Translation for Heterogeneous Data", Verso report, 191, 2001.

....of another XML document. A survey of ve XML languages is presented in [6] XSL [9] is perhaps the most established of these languages, being used in several tools concerned with the presentation of XML into HTML. In common with other approaches to representing XML or semi structured data, e.g. [20, 3, 2, 5, 22, 10, 1], we use a graph based data model in which to represent XML data (the HDM) which supports the notion of unique identi ers for XML elements. One di erence with our approach is how the cardinality and ordering information found in XML documents is represented. Our use of the order and count nodes in ....

....data in one table and studies the performance of using relational databases for querying this data. 23] considers the translation in both directions between XML and the relational model, and commercial tools with more limited capabilities exist which also provide this functionality e.g. Oracle8i. [2] considers the translation between SGML and OO data models, using correspondence rules between constructs expressed in the di erent data models. 19, 5] also consider the translation between SGML and OO data models, in this case via virtual graphs. The approach that we have proposed here is a ....

[Article contains additional citation context not shown here]

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proceedings of ICDT'97, 1997.

....of another XML document. A survey of five XML languages is presented in [5] XSL [8] is perhaps the most established of these languages, being used in several tools concerned with the presentation of XML into HTML. In common with other approaches to representing XML or semi structured data, e.g. [19, 3, 2, 4, 21, 9, 1], we use a graph based data model in which to represent XML data (the HDM) which supports the notion of unique identifiers for XML elements. One difference with our approach is how the ordering information found in XML documents is represented. Our use of the order node in the HDM graph allows ....

....data in one table and studies the performance of using relational databases for querying this data. 22] considers the translation in both directions between XML and the relational model, and commercial tools with more limited capabilities exist which also provide this functionality e.g. Oracle8i. [2] considers the translation between SGML and OO data models, using correspondence rules between constructs expressed in the different data models. 18, 4] also consider the translation between SGML and OO data models, in this case via virtual graphs. The approach that we have proposed here is a ....

[Article contains additional citation context not shown here]

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proceedings of ICDT'97, 1997.

....(XTM) syntax. We implemented the processing model proposed in [6] to construct a Topic Map graph model from an XTM document. RDF is closely related to the concept of semi structured data, identified in the database community [11] 15] as a means for data integration [10] 14] and transformation [4]. Any kind of data that can be represented as a graph is called semi structured data. Thus, if heterogeneous data sources are transformed into a graph representation in some standard representation format, all this data can be queried with the same query infrastructure in the same query. This ....

Serge Abiteboul, Sophie Cluet, and Tova Milo. Correspondence and Translation for Heterogeneous Data. In 6th International Conference on Database Theory, 1997.

....needs to maintain up to tens thousands of links between different document representations. 7 27 2001 7:23 PM 8 Theoretically speaking, the problem of direct translation of the trees encoded with XML is quite complicated and requires a number of significant restrictions to be solvable [Abiteboul et al., 2001]. In addition to high document complexity, a marketplace faces the problem of complex (and exponential in their number) mappings between the objects. An attempt to resolve two difficult problems in one shot naturally fails. We search for a solution to them in a layered approach for B2B ....

Abiteboul, S., Cluet, S., and Milo, T., "Correspondence and Translation for Heterogeneous Data", Verso report, 191, 2001. 7/27/2001 7:23 PM 30

....of product names. Consequently, the next initiative, UCEC 2 provides an extension to the UN SPSC standard with attributes. It 1 www.BizTalk.org 2 www.ucec.org uses only four top levels of the UN SPSC hierarchy and provides a couple of attributes for each category, for example, the category [44 12 15 05] Specialty envelopes has six attributes: type, length, width, weight, colour, and composition. Another horizontal standard, ecl ss, supports the flow of products and information along the supply chain of an industrial enterprise and it is mainly used in Germany. It provides more than 12,700 ....

.... integration framework must be based on current achievements in the ontology integration area [11] Similar problems have been existed in the databases community [13] and a number of schema transformation techniques applicable to the ontology integration problem have been proposed (cf. 14] [15]) Acknowledgement The author would like to thank Dieter Fensel and Michel Klein for their helpful discussions and comments. 1 http: sesame.aidministrator.nl ....

Abiteboul, S., Cluet, S., and Milo, T., "Correspondence and Translation for Heterogeneous Data", Verso report 191, 2001.

....of another XML document. A survey of five XML languages is presented in [5] XSL [7] is perhaps the most established of these languages, being used in several tools concerned with the presentation of XML into HTML. In common with other approaches to representing XML or semi structured data, e.g. [18, 3, 2, 4, 20, 8, 1], we use a graph based data model in which to represent XML data (the HDM) which supports the notion of unique identifiers for XML elements. One difference with our approach is how the ordering information found in XML documents is represented. Our use of the order node in the HDM graph allows ....

....in one table and studies the performance of using relational databases for querying this data. 21] considers the translation in both directions 13 between XML and the relational model, and commercial tools with more limited capabilities exist which also provide this functionality e.g. Oracle8i. [2] considers the translation between SGML and OO data models, using correspondence rules between constructs expressed in the different data models. 17, 4] also consider the translation between SGML and OO data models, in this case via virtual graphs. The approach that we have proposed here is a ....

[Article contains additional citation context not shown here]

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proceedings of ICDT'97, 1997.

....of what distinguish our focus on the constraint mapping problem from other integration efforts and how our approach can be applied in the common mediation architecture [1, 2] There has also been much work on data translation and schema integration. The main focus of these related efforts (e.g. [20, 21, 22, 23, 24, 25]) is to unify data representations across mismatched domains by transforming data to a unified context, where queries can be performed. In contrast, our complementary goal is to map queries to the native domain where data reside. We believe our approach is especially well suited for autonomous ....

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proc. of the 6th ICDT, 1997.

....of what distinguish our focus on the constraint mapping problem from other integration efforts and how our approach can be applied in the common mediation architecture [1, 2] There has also been much work on data translation and schema integration. The main focus of these related efforts (e.g. [20, 21, 22, 23, 24, 25]) is to unify data representations across mismatched domains by transforming data to a unified context, where queries can be performed. In contrast, our complementary goal is to map queries to the native domain where data reside. We believe our approach is especially well suited for autonomous ....

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proceedings of the 6th International Conference on Database Theory, Delphi, Greece, 1997. Springer, Berlin.

....but also we can use conventional database access methods. We suggest a schema extraction method for semistructured data to provide a way of schema based management upon an object oriented data model. The motivation and research trends of semistructured data are found in some recent studies [1,2,3,4,9,10]. The studies mainly focus on representation models and query languages. General introduction and the survey on remarkable data models and query languages are presented in [1] Semistructured data models use similar representation methods, like graph (or tree) structures, to describe data objects ....

....General introduction and the survey on remarkable data models and query languages are presented in [1] Semistructured data models use similar representation methods, like graph (or tree) structures, to describe data objects with their attributes and values. Our study takes the data forest model [2] for the representation of semistructured data. The data forest model describes a semistructured data set as a forest of trees compared to earlier models, like OEM [9] or labeled tree [3,4] that used a single tree (or graph) 2 One recent study [7] tried to infer structural information in ....

S. Abiteboul, Sophie Cluet, and Tova Milo, "Correspondence and Translation for Heterogeneous Data," Proc. '97 ICDT `97, Delphi, Greece, 1997, pp.352-363.

....Our work deals with the problems in information gathering layer. More specifically, we are interested in data mapping problem and we introduce schema construction operators for that problem. The importance and the motivation about schemaless data representations and manipulations are discussed in [1][3] 11] 12] Schemaless data instances are usually described by their attributes and corresponding values. Attribute value pairs were used for data representation in OEM(Object Exchange Model) 11] Labeled Tree[3] and Data Forest Model[1] Although not developed as a semistructured data model, O ....

....data representations and manipulations are discussed in [1] 3] 11] 12] Schemaless data instances are usually described by their attributes and corresponding values. Attribute value pairs were used for data representation in OEM(Object Exchange Model) 11] Labeled Tree[3] and Data Forest Model[1]. Although not developed as a semistructured data model, O 2 s complex value model[8] shows a good way of semistructured representation with attribute value pairs. All the earlier models for semistructured data are similar to each other and their expressive powers are almost same. TSIMMIS 3 ....

Abiteboul, S., Cluet, S., Milo, T.: Correspondence and Translation for Heterogeneous Data. Proceedings of the '97 ICDT, Delphi, Greece (1997) 352--363

....of the translation process: The librarians analyze the source and target contexts (i.e. their schemas) and define the mapping rules, which then drive the constraint mapping algorithm to perform data translation. There are many ways one can go about translating digital library information (e.g. [1, 2, 3, 8, 14, 15], see Section 5) We believe that our approach offers the following advantages over other proposals. ffl Because we view information as constraints, our machinery can provide minimal translations without explicit user guidance. For example, if several rules are available for translating say ....

....g S = ref P = f ref id tr no title authors inst date kws authors name name lname fname date month year kws term g D = all map to string (b) Conceptual schema G ref for SGML Ref documents. id title authors inst date CG98 Constraint . Stanford Univ. kws ref name[1] name[2] lname fname lname fname year month term[1] term[2] Chang Kevin Garcia Molina Hector January 1998 query . data . tr no SIDL . c) Example instance tree I ref of schema G ref . Figure 6: Schema and instance for SGML Ref. or the evolving XML) documents are ....

[Article contains additional citation context not shown here]

Serge Abiteboul, Sophie Cluet, and Tova Milo. Correspondence and translation for heterogeneous data. In Proc. of ICDT '97, Delphi, Greece, 1997. Springer.

....about an information source. Each source may have a wrapper that works as an interface between the mediator and the source itself. TSIMMIS (Garcia Molina et al. 1997) is an example of such mediator based systems. Defining a common data model and then mapping the source and target to it (Abiteboul, Cluet, and Milo 1997) . Specifying a set of matching rules that directly translate between source and target (Milo and Zohar 1998) In general, heterogeneous databases that need to be integrated cover similar domains and their terminology overlaps significantly. The integrators deal with different representations ....

Abiteboul, S., Cluet, S. and Milo, T. 1997. Correspondence and Translation for Heterogeneous Data. In Proceedings of the International Conference on Database Theory (ICDT).

.... Institut f r Wirtschaftsinformatik, Humboldt Universit t zu Berlin myra wiwi.hu berlin.de k Institut f r Informationssysteme, Medizinische Universit t zu L beck linnemann ifis.mu luebeck.de Document transformations and unparsing. The tree transformations discussed in [KP93, FW93, ACM97] can be used to transform the syntax trees of structured documents. These approaches do not extend to the unparsing problem, i.e. the transformation of a possibly complex semantic value into a syntax tree. In [ACM95] an algorithm for the unparsing of database contents is given. The somewhat ....

Serge Abiteboul, Sophie Cluet, and Tova Milo. Correspondence and translation for heterogeneous data. In ICDT'97, 1997. to appear.

....which of type D2. ffl Given two DTD s how can we automate, or semiautomate the translation process The first problem is a type inference problem, and has not been addressed so far in semistructured data (except in a restricted setting in [BDFS97] The second problem has been addressed in [ACM97, MZ98]. Type checking As mentioned before type checking is important in data translation. In a broader sense it can be used to assist the user in query formulation. Given the DTD in Fig. 4, a path expression like Bib.paper.address doesn t make sense (since the tag address is not mentioned in the DTD) ....

Serge Abiteboul, Sophie Cluet, and Tova Milo. Correspondence and translation for heterogeneous data. In ICDT, 1997.

....of the translation process: The librarians analyze the source and target contexts (i.e. their schemas) and define the mapping rules, which then drive the constraint mapping algorithm to perform data translation. There are many ways one can go about translating digital library information [3, 4, 5, 6]. We believe that our approach offers the following advantages over other proposals. ffl Because we view information as constraints, our machinery can provide minimal translations without explicit user guidance. For example, if several rules are available for translating say first, middle and ....

....of modern information systems [1, 8] In this paper we have presented our data translation framework with the goal of coping with semantic or schematic inconsistency in data exchange. Fully automatic semantic integration is extremely difficult [1] if not impossible, Thus, most related efforts [4, 5, 6] advocate, like we do, using human specified rule systems (e.g. datalog rules [7] The common idea behind all these approaches is that a rule essentially specifies how target patterns are to be generated from matching source patterns. Our work differs from other rule translation efforts in ....

Serge Abiteboul, Sophie Cluet, and Tova Milo. Cor10 respondence and translation for heterogeneous data. In Proceedings of ICDT '97, Delphi, Greece, 1997. Springer.

....the document interpreter is to gather and combine information from the data sources, as instructed in the document prescription, and to map it according to the document DTD. 2 The integration of heterogeneous information is done using a tree like model which is very similar to those presented in [4] or [3] As in these approaches we are not using a model that encompasses all the source models. Instead we use a generic and minimalist approach. The data structure we use consists of ordered labeled trees. Each tree node can have one of the following types (akin to an SGML terminology) ELEM, ....

....are represented by 1 As this page is generated internally, copyright is not an issue here. 2 The interpreter itself does not validate the document against the DTD: that can be achieved by coupling an editor with the interpreter. 3 Of course this mapping is not unique; for an alternative see [4] Reuse of linked Documents through Virtual Document Prescriptions 5 element nodes with a label corresponding to their tag name. SGML attributes are represented by an attribute node under the element node. In the example in figure 2, expressions like body.table[#FIRST 1] and i.name are path ....

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proceedings of ICDT97, 1997.

....middleware data model. Then, the data is translated from one representation to another. Finally, the data is exported from its second representation to the target. The heart of W3TRANS is its handling of the second phase, namely translations inside the middleware. For that, the TreeLog language [3, 17] is used. It is a rule based declarative language, with a fixpoint semantics. It differs from other rule based languages in its strong 3 tree awareness: it provides syntactic constructs to search and build complex trees that represent data. A significant advantage of the language is that in most ....

....it provides syntactic constructs to search and build complex trees that represent data. A significant advantage of the language is that in most practical cases, translation, correspondences between data elements, and inverse translations, can all be automatically derived from one TreeLog program [3]. It follows that update propagation form a target format back to the source format can be easily supported. 5 W3QL and W3QS W3QL [14] is a high level query language designed to operate against the WWW. W3QS is a system that provides a W3QL implementation and additional services. W3QL is a ....

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and Translation for Heterogeneous Data. In Proc. ICDT, pages 351-363. 1997.

....as a universal data repository, offering access to sources whose data organization varies from strictly structured databases to almost completely unstructured pages, and everything in between. Consequently, much research has recently focused on data integration and data translation systems [10, 6, 9, 8, 17, 13, 2, 19], whose goals are to allow applications to utilize data from many sources, with possibly widely varying formats. These research efforts have established a common data model of semistructured data, for uniformly representing data from any source. Recently, however, it is being realized that having ....

....and even necessary, in translation and integration systems to support tasks such as query formulation, decomposition and optimization, or declarative specification of data translation. As an example, which we use for motivation throughout the paper, recently suggested tools for data translation [2, 11, 19] use the semistructured data model as a common middleware data model to which data sources are mapped. Translation from source to target formats is achieved by (1) importing data from the source to the middleware model, 2) translating it to another middleware representation that better fits the ....

[Article contains additional citation context not shown here]

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proc. ICDT 97, pages 351--363, 1997.

....data. The reason for this is their lack of flexibility. New models have thus been proposed, integration models in which the notion of type schema disappears or is loosen (see for example [9, 13, 33, 34, 35, 36, 37] Each of these models supports a query language. Some migration tools (e.g. [2, 8, 12, 13]) and optimization techniques are also proposed (e.g. 8, 14] Exporting Data on the Web Most database vendors now provide tools allowing automatic display on the Web of database stored information (see for example O 2 [30] and Oracle [32] These tools generate HTML pages in a rather simple ....

....However, this default mapping is sometimes inadequate and appropriate customization is then either difficult or impossible. The outward simplicity of the Web exportation problem may explain the disinterest of the database research community. To our knowledge, there exist few studies on this topic ([2, 37]) The readers should now be convinced that Web applications raise many interesting problems for database researchers. This paper does not develop all of them but concentrates on the problems of modeling and querying (efficiently) semi structured data. It is by no means exhaustive. It compares ....

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proc. ICDT, Ath`enes, Gr`ece, 1997.

....on the Web in distinct heterogeneous sources, stored under different formats: a specific database vendor format, SGML or Latex (documents) DX formats (scientific data) Step (CAD CAM data) etc. Their integration is a very active field of research, see for instance, for a very small sample, [18, 12, 16, 15, 27, 24, 17, 3, 4]) A key observation is that, often, the application programs used by organizations can only handle data of a specific format. e.g. Web browsers, like Netscape, expect files in HTML format, and relational databases expect relations) To enable a specific tool to manipulate data coming from ....

.... text translator [1] Writing such a program is typically a non trivial task which is often complicated by numerous technical aspects of the specific data sources that are not really relevant to the translation process (e.g. HTML or SGML parsing, or specific database access protocol) Recent works [4, 19] consider a more general framework which enables a more flexible translation between various models. The solution is based on using a common data model to which the source target data is mapped, and providing a common translation language which enables the specification and customization of the ....

[Article contains additional citation context not shown here]

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proc. Int. Conf. on Database Theory (ICDT), 1997.

....heterogeneous data sources very easily. Only minor changes were required for YAT to provide data integration (as opposed to data conversion) in a Web environment. Finaly, we report on our experience while building the Verso Web site using YAT. 1 Introduction Many semistructured data models (e.g. [11, 5, 8, 1]) and appropriate query languages (e.g. 2, 7, 4] have been proposed in the last few years. All of them have been designed as a means to integrate heterogeneous, more or less regular data. We pursued a complementary approach and focused on a system to support data conversion (as opposed to ....

....for the language, some type checking facilities, as well as various mechanisms to combine or compose programs. Its originality, as compared to other similar approaches, can be summarized as follows: YAT vs other semistructured based systems: As previous semistructured data models (e.g. [11, 5, 8, 1]) we rely on a simple representation of graphs to capture data coming from heterogeneous sources. But YAT also provides (i) the possibility to order the outgoing edges from a vertex (in the style of [1] and most importantly (ii) the ability to capture various levels of representation via a ....

[Article contains additional citation context not shown here]

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proc. Int. Conf. on Database Theory (ICDT), Delphi, Greece, January 1997.

....on the design of optimizable languages for the efficient manipulation of large amounts of data. Recently, many languages have been proposed to support the above operations on semistructured data [3] or XML. For instance, Lorel [7] XML QL [22] and XQL [41] have been used for querying, datalog [4] and XSLT [16] for conversion, MSL [38] and the first version of YATL [19] for integration. Yet, due to the number of new features required by each of these tasks and their complex interactions, none of the above languages supports all of them at once. We believe that the key that enabled us to ....

....using nested loop or sort merge joins, etc) We refer the reader to the database literature for more on the subject [39, 29, 26, 37] We would like to stress that the technical material in this paper is not new. Information about the first YATL, following the original datalog approach proposed in [4], can be found in [19, 42] Use of YATL for data conversion and integration is illustrated in [19, 43, 42, 6] More about the new YA T L is still unpublished [44] but its use as a query language is illustrated in [24] Last, the XML algebra underlying YA T L and the corresponding optimization ....

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proceedings of International Conference on Database Theory (ICDT), volume 1186 of Lecture Notes in Computer Science, pages 351--363, Delphi, Greece, Jan. 1997.

....needs to take place. The naive way to translate data between models, is writing a program for each translation task, where this program is suited for the requirements of the specific translation. Examples are Latex to HTML translator [Dra96] or the HTML to text translator [htm96] Recent works [ACM95a, CDSS97] consider a more general environment, which enables a more flexible translation between various data models. The solution is often based on using a common data model to which the source target data is mapped, and providing a common translation language which enables the specification and ....

....For that matter we have defined a common schema and data models in which the schema matching and the data translation take place. We refer to them as the middleware models. The schema model consists of graphs, and the data model consists of labeled forests and is similar to the one introduced in [ACM95a] and to the Object Exchange Model introduced in [PGMW95] The difference is that we allow some nodes to be ordered. This is crucial for modeling data that might be ordered (e.g. structured documents) A system that wishes to use our translation mechanism needs to support an import and export ....

[Article contains additional citation context not shown here]

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proc. of the ACM SIGMOD Conf. on Management of Data, San Jose, California, 1995.

....it) as well as the management of updates of both source and target data will be considered in future works. In short, YAT relies on a data model allowing a rich and uniform representation of data and and a rule based language. The YAT Model: In a manner similar to, e.g. PGMW95, BDS95, FFLS97, ACM97] we use a simple representation of graphs to capture data coming from heterogeneous sources. In contrast to most previous works (and in the style of [ACM97] the outgoing edges from a vertex are ordered. This is essential to capture ordered collections that are frequent in data exchange ....

....rich and uniform representation of data and and a rule based language. The YAT Model: In a manner similar to, e.g. PGMW95, BDS95, FFLS97, ACM97] we use a simple representation of graphs to capture data coming from heterogeneous sources. In contrast to most previous works (and in the style of [ACM97] the outgoing edges from a vertex are ordered. This is essential to capture ordered collections that are frequent in data exchange formats, or in documents. The main originality of the model is its ability to capture various levels of representation. A YAT model can be instantiated into another ....

[Article contains additional citation context not shown here]

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proc. Int. Conf. on Database Theory (ICDT), Delphi, Greece, January 1997.

....as the management of updates of both source and target data will be considered in future works. In short, YAT relies on a data model allowing a rich and 1 For Yet Another Tree based system. uniform representation of data and a rule based language. The YAT Model: In a manner similar to, e.g. [10, 3, 8, 11], we use a simple representation of graphs to capture data coming from heterogeneous sources. In contrast to most previous work (and in the style of [11] the outgoing edges from a vertex are ordered. This is essential to capture ordered collections that are frequent in data exchange formats, or ....

....Another Tree based system. uniform representation of data and a rule based language. The YAT Model: In a manner similar to, e.g. 10, 3, 8, 11] we use a simple representation of graphs to capture data coming from heterogeneous sources. In contrast to most previous work (and in the style of [11]) the outgoing edges from a vertex are ordered. This is essential to capture ordered collections that are frequent in data exchange formats, or in documents. The main originality of the model is its ability to capture various levels of representation. A YAT model can be instantiated into another ....

[Article contains additional citation context not shown here]

S. Abiteboul, S. Cluet, and T. Milo, "Correspondence and translation for heterogeneous data," in Proc. Int. Conf. on Database Theory (ICDT), Delphi, Greece, Jan. 1997.

....information) ii) Then, the data is translated from one abstract representation to the other. iii) Finally, the data is exported to the target. Fig. 1 displays our translation scheme. For part (ii) of the translation scheme (translations inside the middleware) we use the TreeLog language [3, 29]. It is a rule based language, with a fixpoint semantics, in the flavor of Datalog. It differs from other rule based languages in its strong tree awareness: it provides syntactic constructs to search and build complex trees that represent data. For lack of space we cannot give here a formal ....

....build complex trees that represent data. For lack of space we cannot give here a formal definition of the middleware model and the TreeLog language, nor can we demonstrate how the language is used for defining useful translations. More details and examples of translation programs can be found in [3, 29]. Nevertheless, we would like to point out one significant advantage of the language with respect to updates. Assume some data has been extracted and translated from one format to the other, and then was updated in the target system. To propagate the update back to the source, one needs to know ....

[Article contains additional citation context not shown here]

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and Translation for Heterogeneous Data. In Proc. Int. Conf. on Database Theory (ICDT) 1997, pages 351-363. 1997.

No context found.

Abiteboul, S., Cluet, S., and Milo, T. (1997). Correspondence and Translation for Heterogeneous Data. In Intl. Conf. on Database Theory. Delphi, Greece.

No context found.

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and Translation for Heterogeneous Data. In Intl. Conf. on Database Theory, 1997.

No context found.

Abiteboul, S., Cluet, S. & Milo, T. (1997), Correspondence and translation for heterogeneous data, in F. N. Afrati & P. Kolaitis, eds, `Database Theory - ICDT '97, 6th International Conference, Delphi, Greece, January 8-10, 1997, Proceedings', Vol. 1186 of Lecture Notes in Computer Science, Springer, pp. 351--363.

No context found.

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and Translation for Heterogeneous Data. In Intl. Conf. on Database Theory, 1997.

No context found.

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proceedings of 6th Int. Conf. on Database Theory (ICDT), 1997.

No context found.

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proceedings of 6th Int. Conf. on Database Theory (ICDT), 1997.

No context found.

Serge Abiteboul, Sophie Cluet, and Tova Milo. Correspondence and translation for heterogeneous data. In Proceedings of the 6th International Conference on Database Theory, Delphi, Greece, 1997. Springer, Berlin.

No context found.

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In Proceedings of the 6th International Conference on Database Theory, Delphi, Greece, 1997. Springer, Berlin.

No context found.

S. Abiteboul, S. Cluet, and T. Milo. Correspondence and translation for heterogeneous data. In International Conference on Database Theory, January 1997.

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