YANG,H.Z.AND LARSON,P.- A. 1987. Query transformation for PSJ-queries. In Proceedings of 13th International Conference on Very Large Data Bases (Brighton, England, Sept.), pp. 245--254.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... query on the sources would be q (title, subject) V(title, year, author) V2(title, subject) The LAV reformulation problem has received significant attention because of its relevance to other database problems, such as query optimization [7] maintaining physical data independence [22, 19], and data warehouse design. The GAV and the LAV strategies can be qualitatively or quantitatively [6] compared in terms of their adequacy (a) to model a particular integration situation, b) to cope with autonomy of the sources (sources changing their exported schemas, joining or leaving the ....

Yang, H. Z., and Larson, P. A., 1987. Query Transformation for PSJ Queries. Proceedings of the International Conference on Very Large Data Bases (VLDB), Brighton, England, pages 245-254.

....practical bene t from all these restrictions. In the end, the exponential behavior of the GMAP algorithm and the diculties we had to resolve for the backchase phase are closely related. Our experiments include schemas, views and queries of signi cantly bigger complexity than those reported in [22, 20, 5]. Their experiments show that using views can be done and in the case of [20] that it can produce faster plans. But [22] measures only optimization time and [20] does not separate the cost of the optimization itself, so they do not o er any numbers that we can compare with our time reduction ....

....had to resolve for the backchase phase are closely related. Our experiments include schemas, views and queries of signi cantly bigger complexity than those reported in [22, 20, 5] Their experiments show that using views can be done and in the case of [20] that it can produce faster plans. But [22] measures only optimization time and [20] does not separate the cost of the optimization itself, so they do not o er any numbers that we can compare with our time reduction gures (section 5.4) 5] shows a very good behavior of the optimization time as a function of plans produced, but cannot be ....

H.Z. Yang and P.A. Larson. Query transformation for psj queries. In Proceedings of the 13th International VLDB Conference, pages 245-254, 1987. 12

....The experimental configurations cover both relational and OO optimization, and the scenarios considered exhibit the systematic interaction between semantic and physical optimization that allows us to find high quality plans. Our experiments cover and go beyond the experiments of [CGK 99, TSI94, YL87, SO89] We reconstructed those experiments and found that our optimizer can also find the desired plans for a set of chosen queries. However, we went further by repeating the experiments on families of queries and schemas of similar structure but of increasing complexity. This allows us to find ....

....the same way we rewrite queries into plans that use indexes. Materialized views Source capabilities Materialized conjunctive or PSJ (project selectjoin) views, or cached results of conjunctive PSJ queries over a relational schema R have been used in answering other conjunctive PSJ queries over R [YL87, CR94, CKPS95, LMSS95, Qia96] We consider the more general form Here we denote by P x an arbitrary sequence of bindings P 1 x 1 ; Pn xn , by O( x) we denote the fact that variables x 1 ; xn can appear in the output record O (and similar for B( x) Like indexes, such ....

[Article contains additional citation context not shown here]

H.Z. Yang and P.A. Larson. Query transformation for psj queries. In Proceedings of the 13th International VLDB Conference, pages 245--254, 1987.

....of the logical schema from the physical implementation, while in information integration systems most difficulties come from heterogeneity. There have been several research efforts investigating physical data independence as the central issue [45, 20] or investigating closely related problems [48, 16, 27, 15, 30, 39, 38]. All of them recognize physical data independence as an optimization problem: rewrite a query Q( written against a logical schema into an equivalent query plan Q ( Phi) written against a physical schema Phi, given a semantic relationship between and Phi. The question is how to define, ....

....use indexes. We leave the details out due to lack of space. Materialized views Source capabilities Materialized conjunctive or PSJ (project select join) views, or cached results of conjunctive PSJ queries over a relational schema R have been used in answering other conjunctive PSJ queries over R [48, 16, 15, 30, 38]. We consider the more general form = select O( x) from P x where B( x) Here we denote by P x an arbitrary sequence of bindings P 1 x 1 ; Pn xn , by O( x) we denote the fact that variables x 1 ; xn can appear in the output record O (and similar for B( x) Like indexes, ....

H.Z. Yang and P.A. Larson. Query transformation for psj queries. In Proceedings of the 13th International VLDB Conference, pages 245--254, 1987.

....data may be constantly added removed from the view. Yet, we believe that there has been no prior formal modeling of the evolution process. 4 Related research involves the containment property of a materialized view with respect to its base data: a few of the many references in this area include [25, 6, 16, 2, 13]. However, the temporal aspects of content evolution have not been systematically addressed in this work. Refresh policies for materialized views have been previously discussed in the literature (e.g. 17] and [8] Typically, materialized views are refreshed immediately upon updates to the base ....

H.Z. Yang and P.A. Larson. Query transformation for PSJ-queries. In Proceedings of the 13th International VLDB Conference, pages 245--254, 1987.

....data may be constantly added removed from the view. Yet, we believe that there has been no prior formal modeling of the evolution process. 3 Related research involves the containment property of a materialized view with respect to its base data: a few of the many references in this area include [38, 6, 19, 2, 13]. However, the temporal aspects of content evolution have not been systematically addressed in this work. In [2] for example, the containment relationships between a materialized view I and the true query result V(D) taken from a database D, can be either I = V(D) or I V(D) 1 ....

H.Z. Yang and P.A. Larson. Query transformation for PSJ-queries. In Proceedings of the 13th International VLDB Conference, pages 245--254, 1987.

....we obtain are union queries. Our attention is focused on how to quickly find rewritings that give as many correct answers as possible, rather than on how e#ciently the rewritings can be evaluated. 1. 1 Previous Work Apparently the first papers dealing with query rewriting using views are [3] and [4], in which equivalent rewritings of conjunctive queries are discussed. Over the past few years, the problem has received intensive attention mainly because of its relevance to data integration and query optimization. For a comprehensive survey, see [5] Here we only mention a few papers that are ....

H. Z. Yang and P.-A. Larson. Query transformation for PSJ-queries. In VLDB, pages 245--254, 1987.

....1 Introduction The problem of answering queries using views (a.k.a. rewriting queries using views) has recently received significant attention because of its relevance to a wide variety of data management problems [20] query optimization [6, 21, 36] maintenance of physical data independence [35, 33, 27], data integration [22, 9, 18, 19] and data warehouse and web site design [16, 32] Informally speaking, the problem is the following. Suppose we are given a query over a database schema, and a set of view definitions 4670 38 over the same schema. Is it possible to answer the query ....

....to answer the query using only the answers to the views 37536 , and if so, how There are two main contexts in which the problem of answering queries using views has been considered. In the first context, where the goal is query optimization or maintenance of physical data independence [35, 33, 6], we search for an expression that uses the views and is equivalent to the original query. Here it is usually assumed that the number of views is on the same order as the size of the schema. The second context is that of data integration, where views describe a set of autonomous heterogenous data ....

[Article contains additional citation context not shown here]

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In VLDB, 1987.

....are similar to views defined on them. The problem of 4 The period of 1950 present implies the period 1950 1989, since the data presented here was published by Gary T. Whiteford in 1989. creating an execution plan is thus closely related to the problem of answering queries using views [YL87, LMSS95, CKPS95]. It was shown to be NP complete in [LMSS95] The algorithm we describe here uses domain and source characteristics and is therefore tractable and efficient. 3.1 Operator nodes used in the plan Following are the different types of operator nodes used in the plans. We use the term result set ....

H. A. Yang, and P. A. Larson. Query transformation for PSJ queries. Proceedings of 13th International Conference on Very Large Databases VLDB87, Brighton, England, 1987, pp. 245-254.

....to refer to row IDs. 1 This independence enables us to modify the storage schema of the data (i.e. the physical view) without changing its logical schema, and to model more complex types of indices. Hence, several authors describe the storage schema as a set of views over the logical schema [YL87, TSI96, Flo96] Given these descriptions of the storage, the problem of computing a query execution plan (which, of course, must access the physical storage) involves guring out how to use the views to answer the query. A second class of applications in which our problem arises is data ....

....but also to make a judicious cost based decision on when to use the available views. 2. 2 Maintaining Physical Data Independence Several of the works on answering queries using views were inspired by the goal of maintaining physical data independence in relational and object oriented databases [YL87, TSI96, Flo96] One of the principles underlying modern database systems is the separation between the logical view of 4 Evaluation Advises Quarter Quarter Faculty Department Course name area name c number worksIn teaches title major name registered Student Figure 1: An ....

[Article contains additional citation context not shown here]

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In Proc. of the Int. Conf. on Very Large Data Bases (VLDB), pages 245-254, Brighton, England, 1987. 42

....of the physical view of the data and its logical view. This independence enables us to modify the storage schema of the data (i.e. the physical view) without changing its logical schema. Hence, several authors have proposed to describe the storage schema as a set of views over the logical schema [45, 43, 24, 19]. Given these descriptions of the storage, the problem of computing a query execution plan (which, of course, must access the physical storage) involves guring out how to use the views to answer the query. A second context in which our problem arises is data integration. A data integration ....

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In Proc. of VLDB, pages 245-254, Brighton, England, 1987.

....logical view. This independence enables us to modify the storage schema of the data (i.e. the physical view) without changing its logical schema, and to model more complex types of indices. Hence, several authors proposed to describe the storage schema as a set of views over the logical schema [YL87, TSI96, Flo96] Given these descriptions of the storage, the problem of computing a query execution plan (which, of course, must access the physical storage) involves figuring out how to use the views to answer the query. A second class of applications in which our problem arises is data ....

....teaches title Figure 1: An Entity Relationship diagram for the university domain. 2. 2 Maintaining Physical Data Independence Several of the works on answering queries using views were inspired by the goal of maintaining physical data independence in relational and object oriented databases [YL87, TSI96, Flo96] One of the principles underlying modern database systems is the separation between the logical view of the data (e.g. as tables with their named attributes) and the physical view of the data (i.e. how it is layed out on disk) Relational database systems, even though they ....

[Article contains additional citation context not shown here]

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In Proc. of the Int. Conf. on Very Large Data Bases (VLDB), pages 245--254, Brighton, England, 1987.

.... approaches to information integration, the global as view and the local asview approach 1 [10] The mediation process performed by the former is guaranteed to run in polynomial time, while the reasoning problem of the localas view approach is presumably very hard even for simple query languages [32, 19, 30]. On the other hand, the latter is more flexible in dealing with mappings of sources, allows to add new sources more easily, and has better completeness properties w.r.t. query evaluation. Also with respect to ease of use and extensibility, the local as view approach is usually considered ....

....approach is the one of answering queries using logical views. It basically decomposes into the task of finding a rewriting s of an input query q in terms of the sources and checking query containment s q, that is, that every result tuple of s is also a result of q, for any possible database [32]. These two problems are individually NP complete already in the simple case of conjunctive queries [7, 19] i.e. queries that support selections, projections and joins, the basic operations of relational algebra) however, the combined problem remains in NP [19] Algorithms for solving the query ....

H. Z. Yang and Per- Ake Larson. Query Transformation for PSJ-Queries. Proc. 13th VLDB 1987, Brighton, England, pp. 245--254. 12

....we need only look up the Price and Finance attribute from Dealer3 using the value of CarID for each tuple in the current answer set. This is likely to be much cheaper than computing a three way join from scratch. 2 1.2. Related Work Previous work on using views to answer queries includes [LY85, YL87, SJGP90, TSI94, CR94, CKPS95, LMSS95, RSU95] Subsequently, GHQ95, SDJY96] discuss how to answer aggregate queries using materialized aggregate views. TSI94, CKPS95] also describe how to incorporate their techniques in a traditional cost based query optimizer. The problem of view adaptation is ....

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In VLDB, pages 245--254, 1987.

....to the queries describing the sources. The problem of answering a query using a set of previously materialized views has received signi cant attention because of its relevance to other database problems, such as query optimization Chaudhuri et al. 1995, maintaining physical data independence Yang and Larson, 1987; Tsatalos et al. 1996, and data warehouse design. Formally, suppose we are given a query Q and a set of view de nitions V 1 ; Vm . A rewriting of the query using the views is a query expression Q 0 whose subgoals use only view predicates or interpreted predicates. We Logic Based ....

Yang, H. Z. and Larson, P. A. (1987). Query transformation for PSJqueries. In Proc. of the Int. Conf. on Very Large Data Bases (VLDB), pages 245-254, Brighton, England.

....source schemas. In this case, reformulation amounts to unfolding the user s query. Local as view (LAV) 82, 72, 37, 38, 51, 73] the data sources are described as queries over the relations in the mediated schema. Here query reformulation reduces to the problem of answering queries using views [80, 36, 123, 117, 23, 108]. Description Logics: 22, 82] the mediated schema and the data sources are described as a terminology in some Description Logic. Query reformulation makes use of the subsumption and satis ability algorithms provided by the Description Logic system. Planning operators: 40, 54, 7, 72] data ....

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In Proc. of the Int. Conf. on Very Large Data Bases (VLDB), pages 245-254, Brighton, England, 1987.

....visiting Stanford University. 1 Introduction The notion of materialized view is essential in databases [Ull88] and is attracting more and more attention with the popularity of data warehouses [LZW 97] The problem of answering queries using materialized views [LMSS95, CKPS95, DG97a, BLR97, YL87, RSU95, LRU96, Ull97, DL97, Dus97, DG97b, LRO96] has been studied intensively. We propose a systematic study of its complexity. We also briefly consider the related problems of view consistency and view self maintainability [Huy97] Our results exhibit strong connections with two among the most ....

H. Z. Yang, P.- A. Larson. Query transformation for PSJ-queries. In VLDB-87, pp. 245--254. 10

.... to information integration, the global as view and the local as view approach 1 [10] The mediation process performed by the former is guaranteed to run in polynomial time, while the reasoning problem of the local as view approach is presumably very hard even for simple query languages [32, 19, 30]. On the other hand, the latter is more flexible in dealing with mappings of sources, allows to add new sources more easily, and has better completeness properties w.r.t. query evaluation. Also with respect to ease of use and extensibility, the local as view approach is usually considered ....

....approach is the one of answering queries using logical views. It basically decomposes into the task of finding a rewriting s of an input query q in terms of the sources and checking query containment s # q, that is, that every result tuple of s is also a result of q, for any possible database [32]. These two problems are individually NP complete already in the simple case of conjunctive queries [7, 19] i.e. queries that support selections, projections and joins, the basic operations of relational algebra) however, the combined problem remains in NP [19] Algorithms for solving the query ....

H. Z. Yang and Per- Ake Larson. Query Transformation for PSJ-Queries. Proc. 13th VLDB 1987, Brighton, England, pp. 245--254. 12

....1 ) AG(Q 2 ) and SG(MV 3 ) AG(Q 2 ) as shown in Figure 8 (b) two subquery blocks for MV 1 and MV 3 can be integrated into a UNIONALL GROUP BY query, resulting in Q 2 in Figure 4. 5. Related Work Several work on answering conjunctive queries using MVs has been proposed in the literature [YL87, CR94, LMSD95, CKPS95]. LMSD95] formalized the problem of finding rewritings of a conjunctive query under set semantics in terms of containment mappings from views to the query. CKPS95] addressed the problem of optimizing queries using MVs. They proposed a rewriting method for conjunctive SPJ queries and integrated ....

H. Z. Yang and P. A. Larson. Query Transformation for PSJ-Queries. In Proc. of 13th Int'l Conf. on Very Large Data Bases, Brighton, pp.245--254, August 1987.

....to be careful with the converse statement that a function is just a view, so almost anything you can think of is a view. 6 logically accessing large collections is an indispensable technology for Web applications. We should mention also more recent results on answering queries using views, e.g. [12, 20, 34, 4, 25], often developed already with Web applications in mind. Tag line XML Views require standard database view technology, but much more. Tag line What databases can bring to XML is query optimization and query rewriting. Issues Query language for XML. In particular, the management of links and ....

H.Z. Yang and P.- A. Larson. Query transformation for PSJ-queries. In Proc. Int. Conf. on Very Large Data Bases (VLDB), 1987. 18

....1 Introduction The problem of answering queries using views (a.k.a. rewriting queries using views) has recently received significant attention because of its relevance to a wide variety of data management problems [20] query optimization [6, 21, 36] maintenance of physical data independence [35, 33, 27], data integration [22, 9, 18, 19] and data warehouse and web site design [16, 32] Informally speaking, the problem is the following. Suppose we are given a query Q over a database schema, and a set of view definitions V 1 , Vn over the same schema. Is it possible to answer the query Q ....

....answer the query Q using only the answers to the views V 1 , Vn , and if so, how There are two main contexts in which the problem of answering queries using views has been considered. In the first context, where the goal is query optimization or maintenance of physical data independence [35, 33, 6], we search for an expression that uses the views and is equivalent to the original query. Here it is usually assumed that the number of views is on the same order as the size of the schema. The second context is that of data integration, where views describe a set of autonomous heterogenous data ....

[Article contains additional citation context not shown here]

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In VLDB, 1987.

....in the presence of materialized cached views, as a subroutine. By virtue of working on a general DAG structure, our techniques are extensible, unlike the solutions of [CKPS95] and [CR94] The problem of detecting whether an expression can be used to compute another has also been studied in [LY85, YL87, Sel88a] however, they do not address the problem of choosing what to materialize, or the problem of finding the best query plans in a cost based fashion. Our algorithms have been described in the context of a Volcano like optimizer; at least two commercial database systems, from Microsoft and ....

H. Z. Yang and P. A. Larson. Query transformation for psj queries. In Intl. Conf. Very Large Databases, pages 245--254, Brighton, August 1987.

....a mapping, completing a mapping, and determining equivalence of models. For example, a mapping can be thought of as a view of one model in terms of another. Therefore, inverting a mapping resembles the problem of inverting views, which raises the relevance of work on answering queries using views [YL87, TSI96, CKPS95, LMSS95, DG97, FRV96] Even in the relational case these works need to be extended for the problems faced in model management, not to mention extensions to other contexts. Description Logics provide another formalism that has been shown to be useful for modeling DB schemas and ....

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In Proc. of the Int. Conf. on Very Large Data Bases (VLDB), pages 245--254, Brighton, England, 1987.

....classifying the concepts in a concept hierarchy, and then computing them starting with the parent concepts. This corresponds to evaluating a new Classic query (the new concept) using information in several materialized views (the old concepts) Classic has been used for data archaeology. LY85, YL87] look at the question of answering queries using cached results or materialized views. LY85, YL87] show how to transform an SPJ (select project join) query so that it is expressed completely using a given set of views, without any reference to the base relations. They also have the idea of ....

....concepts. This corresponds to evaluating a new Classic query (the new concept) using information in several materialized views (the old concepts) Classic has been used for data archaeology. LY85, YL87] look at the question of answering queries using cached results or materialized views. LY85, YL87] show how to transform an SPJ (select project join) query so that it is expressed completely using a given set of views, without any reference to the base relations. They also have the idea of augmented views where each view is extended with keys of the underlying base relations. CKPS95] tackle ....

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In VLDB, pages 245--254, 1987. A Tables of Adaptation Techniques In this section we present the complete tables of adaptation techniques. The initial view for each table is described in the corresponding section of the text (Section 3.5 for

....confirms that this goal has been met. Generalizations The enumeration algorithm ExOptPlan is robust in that it is completely independent of the algorithm used to generate MapTable. This would make it possible to pick an algorithm for generating equivalent queries using other algorithms [Fin82, YL87, CR94] see discussion in the following section) The optimization algorithm presented in this paper extends to the case where the query and the materialized views are single block Select Project Join queries (i.e. not necessarily conjunctive queries) Most commercial optimize multiblock SQL ....

....the approaches has to be predetermined. Thus, the optimizer can not explore both the options depending on the query and cost estimations. The task of generating equivalent queries based on existing query fragments or semantic knowledge has been studied in several different contexts [Fin82, LY85, YL87, Sel88, CGM90, CS93, CR94] However, all these techniques generate equivalent queries explicitly . In contrast, much of our efficiency in optimization stems from the implicit encoding of the set of equivalent queries in MapTable and a join enumeration algorithm that exploits the encoding. 7 ....

H. Z. Yang and P. A. Larson. Query transformation for psj-queries. In Proceedings of the 13th International VLDB Conference, pages 245--254, Brighton, August 1987.

....site. Client queries can be answered more efficiently if they can be folded using the cached data [9, 6] In the case of a disconnected network, a query can still be answered, at least partially, if it can be folded using views and queries maintained at available sites. A number of algorithms [3, 10, 15, 16, 18] have been developed for folding of conjunctive queries and resources. Some of these, including [3, 16, 18] were developed for purposes of optimization and, accordingly, seek to produce foldings (as in Example 1) that are equivalent to the original query. However, there are also applications for ....

....case of a disconnected network, a query can still be answered, at least partially, if it can be folded using views and queries maintained at available sites. A number of algorithms [3, 10, 15, 16, 18] have been developed for folding of conjunctive queries and resources. Some of these, including [3, 16, 18], were developed for purposes of optimization and, accordingly, seek to produce foldings (as in Example 1) that are equivalent to the original query. However, there are also applications for foldings that are known only to be contained in the original query. For example, a user may wish to ....

H. Z. Yang and P.- A. Larson. Query transformation for PSJ-queries. In Proceedings of 13th VLDB, pages 245--254, 1987.

....is treated in Section 5. Section 6 contains concluding remarks and possible extensions. An extended version of this work appears on [29] 2 Related work The DW configuration problem relates to several overlapping research areas. Answering queries using materialized views has been studied in [14, 34, 30, 15, 4]. The same issue, in connection to aggregate queries and views, has been studied in [7] while in [5] multiset semantics is additionally considered. The problem of optimizing query evaluation in the presence of materialized views has been studied in [30, 4] The problem of maintaining ....

H. Yang and P.-A. Larson. Query Transformation for PSJ-queries. In Proc. of VLDB, pages 245--254, 1987.

....check their compatibility and reject unsound solutions. 6 Related Work Algorithms for rewriting queries have been recently used to satisfy the various needs of several information integration systems [3] 4] and [11] Query rewriting has also been studied for its use in query optimization [12] [13]. Other approaches to query rewriting is using query folding [14] and the inverse rules algorithm [15] The inverse rules algorithm works for recursive queries but generates even more candidate solutions, that need to be verified, than the bucket al..gorithm. Rewriting queries utliizing views with ....

Yang H.Z. and Larson P.A. Query transformation for psj-queries. In Proc. of the 13th International VLDB Conference, pages 254--254, 1987.

....in the presence of materialized cached views, as a subroutine. By virtue of working on a general DAG structure, our techniques are extensible, unlike the solutions of [CKPS95] and [CR94] The problem of detecting whether an expression can be used to compute another has also been studied in [LY85, YL87, Sel88a] however, they do not address the problem of choosing what to materialize, or the problem of finding the best query plans in a cost based fashion. Recently, RR98] considers the problem of detecting invariant parts of a nested subquery, and teaching the optimizer to choose a plan that ....

H. Z. Yang and P. A. Larson. Query transformation for psj queries. In Intl. Conf. Very Large Databases, pages 245--254, Brighton, August 1987.

....conditions that utilize information about set valued queries and views, such as the existence of keys or SELECT DISTINCT. 1 In particular, we assume that there are no view tables in the FROM clause of the query or the view. 2 There has been previous work on using views to answer queries (e.g. [YL87, CKPS95, SJGP90, TSI94, CR94]) but the formal aspects of finding the equivalent rewritings for SQL queries with multiset semantics, grouping and aggregation, have received little attention. Most of the work has focused on conjunctive queries (i.e. without grouping and aggregation) with set semantics. One exception is the ....

....V 1 is: Q 0 : SELECT A1 FROM V1(A1 ; A4 ) WHERE A1 = A4 Observe that, in the absence of the information about keys, Q 0 is not a valid rewriting of Q, and furthermore V is not usable in evaluating Q. 2 6 Related Work There has been previous work on using views to answer queries (e.g. [CKPS95, YL87, SJGP90, TSI94, CR94]) but the formal aspects of this problem, particularly for SQL, have received less attention. 17 Caching of previous query results was explored in [Sel88, SJGP90] as a means of supporting stored procedures. This corresponds to using materialized views when they match syntactically a ....

[Article contains additional citation context not shown here]

H. Z. Yang and P.-A. Larson. Query transformation for PSJ-queries. In Proceedings of the International Conference on Very Large Databases, pages 245--254, 1987.

....by McGraw Hill . 3. Join the results of q1 and q2 to get the answer to the user query. Note that the feasibility of this plan is dependent on the source capabilities. Several works have considered the problem of implementing a query processor that uses the results of materialized views [56,43,46, 12, 55, 10]. 10,55,56]developed solutions to query folding or rewritings for conjunctive queries and resources with built in predicates. These solutions are based on either finding syntactic or 1 1 mappings from the view to the query. Their algorithms use exhaustive search strategies that are ....

....3. Join the results of q1 and q2 to get the answer to the user query. Note that the feasibility of this plan is dependent on the source capabilities. Several works have considered the problem of implementing a query processor that uses the results of materialized views [56,43,46, 12, 55, 10] [10,55,56]developed solutions to query folding or rewritings for conjunctive queries and resources with built in predicates. These solutions are based on either finding syntactic or 1 1 mappings from the view to the query. Their algorithms use exhaustive search strategies that are exponential time in ....

H.Z. Yang and P.A. Larson. Query transformation for psj-queries. In Proc. of Int. Conference on Very Large Databases, pages 245--254, 1987.

.... Review(title; review) The reformulated query on the sources would be: q 0 (title; review) GammaV 1 (title; year; director) V 2 (title; review) The LAV reformulation problem is very closely related to the problem of answering queries using views, studied in the database literature [YL87, TSI96, LMSS95, CKPS95, RSU95, DG97b] This problem has received significant attention because of its relevance to other database problems, such as query optimization [CKPS95] maintaining physical data independence [YL87, TSI96] and data warehouse design. As it turns out, the query ....

.... of answering queries using views, studied in the database literature [YL87, TSI96, LMSS95, CKPS95, RSU95, DG97b] This problem has received significant attention because of its relevance to other database problems, such as query optimization [CKPS95] maintaining physical data independence [YL87, TSI96] and data warehouse design. As it turns out, the query reformulation problem is in general NP complete in the size of the source descriptions and user query even when the queries describing the sources and the user query are conjunctive and don t contain interpreted predicates [LMSS95] ....

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In Proc. of the Int. Conf. on Very Large Data Bases (VLDB), pages 245--254, Brighton, England, 1987.

....Recently, Qian [18] independently showed that acyclic queries admit polynomial time algorithms for containment and minimization. Our work treats acyclic queries as a special case of queries with width 1, and so Qian s algorithm for query containment falls out as a special case. Yang and Larson [14, 22] considered the problem of finding rewritings for SPJ queries using SPJ views. In their analysis, they considered what amounts to 1 1 mappings from the views to the query, and therefore their algorithm can miss some rewritings. The problem of finding equivalent rewritings was studied formally by ....

H.Z. Yang and P.A. Larson. Query transformation for PSJ-queries. In Proceedings of the Thirteenth International Conference on Very Large Data Bases, pages 245--254, 1987.

.... in mediator systems (e.g. Information Manifold [LRO96] TSIMMIS [CGMH 94, PGGMU95] SIMS [AKS96] HERMES [ACPS96] Razor [FW97] Infomaster [DG97] Mobile Computing [BI94, HSW94] view adaptation [GMR95] maintaining physical data independence [TSI96] and speeding up query processing [YL87, CKPS95] The problem arises naturally in mediator systems that provide access to multiple heterogeneous information sources [Ull97] In such systems, The work was done while the first author was at AT T Laboratories. The work of the second and third authors was supported by NSF grant ....

....is solved by a program that uses the views to obtain answers from the information sources. For the solution to be correct, it must produce the same answer as the query (or a maximal subset of it) for any set of tuples in the base relations. Previous work on answering queries using views (e.g. YL87, LMSS95, RSU95, CKPS95, Qia96] assumes the set V of views to be finite. Papakonstantinou et al. PGGMU95] considers the possibility of an infinite set of views supported by one source and shows that infinite sets of views naturally model sources with the ability to answer complex queries about ....

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In Proceedings of the 13th International VLDB Conference, pages 245--254, 1987.

....plan is the same as the problem of rewriting queries using views. In this context, the views are the relations in the sources. The problem of rewriting queries using views has also been investigated in the database literature because of its importance for query optimization and data warehousing [32,30,6,22,26,25,10]. Most previous work has considered the problem of finding query plans where the query plan is required to be equivalent to the original query. In practice, the collection of available data sources may not contain all the information needed to answer a query, and therefore, we need to resort to ....

....by the ability of a source to answer a potentially infinite number of conjunctive queries. Hence, 24] considered how to answer queries given an infinite number of conjunctive source descriptions. Several authors have considered the problem of rewriting queries using views for query optimization [32,6,30]. In this context, one usually requires a query plan that is equivalent to the original query. The algorithms described in [6,30] also explain how to combine the search for query plans with a traditional System R style query optimizer. Another use of rewriting queries using views is explored in ....

H. Z. Yang and P.- A. Larson. Query transformation for PSJ-queries. In Proceedings of the Thirteenth International Conference on Very Large Data Bases, pages 245--254, Los Altos, CA, 1987.

....the focus is on materializing data cubes (aggregates) 14] 1] compared the cost of queries and updates of performing a join using a materialized join with other approaches; 15] studied the performance of partially materialized views. Translating queries to use materialized views was studied in [26, 17, 20, 3, 19]. And maintenance of materialized views was also studied (see [11] Our work is also related to [25] where database schema design also considers access patterns. Our work is also related to [10] where queries to be sped up are key based selections. The paper is organized as follows. We give a few ....

H. Z. Yang and P.- A. Larson. Query transformation for PSJ-queries. In Proc. Int. Conf. Very Large Data Bases, 1987.

....classifying the concepts in a concept hierarchy, and then computing them starting with the parent concepts. This corresponds to evaluating a new Classic query (the new concept) using information in several materialized views (the old concepts) Classic has been used for data archaeology. LY85, YL87] look at the question of answering queries using cached results or materialized views. LY85, YL87] show how to transform an SPJ (selectproject join) query so that it is expressed completely using a given set of views, without any reference to the base relations. They also have the idea of ....

....concepts. This corresponds to evaluating a new Classic query (the new concept) using information in several materialized views (the old concepts) Classic has been used for data archaeology. LY85, YL87] look at the question of answering queries using cached results or materialized views. LY85, YL87] show how to transform an SPJ (selectproject join) query so that it is expressed completely using a given set of views, without any reference to the base relations. They also have the idea of augmented views where each view is extended with keys of the underlying base relations. CKPS95] tackle ....

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In VLDB, pages 245--254, 1987.

....conflict with such security constraints, thereby abiding semantic data security. 15 Such techniques include semantic query optimization (see Section 5.1) semantic query caching ( 23] 15] and [57] cooperative answering (see Section 5. 2) and query rewriting using views ( 81] 65] 14] and [90]) December 1997 ICs: Semantics and Applications Godfrey, Grant, Gryz, Minker p. 25 The issues of semantic data security can be epistemically more complex, if one also wants in ensure that a user can never infer a secret , and thereby compromise database security. In [5] first order ....

H. Z. Yang and P.- A. Larson. Query transformation for PSJ-queries. In Proceedings of the Thirteenth International Conference on Very Large Data Bases, pages 245--254, 1987.

....the memoized functions (i.e. computing and refreshing their values) To goal of the reformulation step is to find a reformulation of the site schema that minimizes the cost function. The problem of reformulating a site schema is closely related to the problem of rewriting queries using views [18, 3, 12, 15, 6]. The major differences between previous works and the problem we solve here are: 1) we need to consider queries with outerjoins, 2) memoized functions can only be used in restricted ways in the rewritings, and (3) we need to also consider the possibility of exploiting known preconditions on the ....

....to materialize and functions to memoize, we need to reformulate the queries in the site definition to use the views and functions. Broadly speaking, the reformulation step can be considered as an instance of the problem of rewriting queries using views, which has been considered in several works [18, 3, 12, 15, 16, 6]. However, there are several issues in our context that introduce additional difficulties and provide opportunities for special optimizations. In this section we describe the differences and and describe an algorithm for reformulation that is especially appropriate in this context. 5.1 ....

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In Proceedings of the 13th International VLDB Conference, pages 245--254, 1987.

....Queries: Generation of MapTable In this section, we discuss how we can use one level rewrite rules to derive queries that are equivalent to the given query in the presence of materialized views. This problem of generating equivalent queries in the presence of views has been studied before [YL87] also see Section 6) In addition to a simplified exposition of the problem for conjunctive queries, the novelty here is in generating an implicit representation of equivalent queries in such a way that the join enumeration phase can exploit it. Our notion of equivalence of queries is as in SQL ....

....3 confirms that this goal has been met. Generalizations The enumeration algorithm ExOptPlan is robust in that it is completely independent of the algorithm used to generate MapTable. This would make it possible to pick an algorithm for generating equivalent queries using other algorithms [Fin82, YL87, CR94] see discussion in the related work) We have also studied several important extensions to the algorithm to efficiently handle nested views and to avoid applications of one level rules where unnecessary. These are described in the full paper [CKPS94] The optimization algorithm presented ....

[Article contains additional citation context not shown here]

H. Z. Yang and P. A. Larson. Query transformation for psj-queries. In Proceedings of the 13th International VLDB Conference, pages 245--254, Brighton, August 1987.

....implication problem P 0 ) PLHS must be solved, where P 0 is the conjunction of all the selection, join, exists and not exists conditions. Unfortunately, there are no known complete algorithms to solve the general implication problem with a mixture of existential and universal quantifiers ([YL87]) In Section 4.3, we develop an algorithm to compute the closure of a conjunctive condition which may include exists conditions but only atomic not exists conditions. This algorithm can be useful in solving a more general implication problem than the one in Line 2. However, we do not show it here ....

H. Yang and P.-A. Larson. Query transformation for PSJ-queries. In Proceedings of the 1987 International Conference On Very Large Data Bases, pages 245--254, October 1987.

No context found.

YANG,H.Z.AND LARSON,P.- A. 1987. Query transformation for PSJ-queries. In Proceedings of 13th International Conference on Very Large Data Bases (Brighton, England, Sept.), pp. 245--254.

No context found.

H. Z. Yang and P.-A. Larson. Query transformation for PSJ-queries. In Proc. VLDB, 1987.

No context found.

Yang H. Z. and P. A. Larson. Query Transformation for PSJ Queries. In Proc. 13th Int'l. Conf. on Very Large Data Bases (VLDB '87) Brighton, England, pp. 245--254. 7

No context found.

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In VLDB, 1987.

No context found.

H. Z. Yang and P.- A. Larson. Query transformation for PSJ-queries. In Proceedings of the Thirteenth International Conference on Very Large Data Bases, pages 245--254, Los Altos, CA, 1987.

No context found.

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In Proc. of VLDB, pages 245-254, Brighton, England, 1987.

No context found.

H. Z. Yang and P. A. Larson. Query transformation for PSJ-queries. In Proc. of VLDB, pages 245-254, Brighton, England, 1987.

No context found.

H. Z. Yang and P. A. Larson. Query transformation for PSJqueries. In Proc. of the Int. Conf. on Very Large Data Bases (VLDB), pages 245--254, Brighton, England, 1987. 75

No context found.

Yang HZ, Larson PA (1987) Query transformation for PSJ-queries. In: Proc Int VLDB Conf, pp 245--254, Brighton, UK

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