E. Wong and K. Yousse . Decomposition-a strategy for query processing. ACM Trans. on Database Systems, 1(3):223-241, 1976.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....constraint logic programming (CLP) systems etc. Traditionally, there have been two major approaches to query optimization. One is based on compile time algebraic simplification of a query using heuristics as in (Hall, 1976, Minker, 1978, Pecherer, 1975, Smith and Chang, 1975, Palermo, 1974, Wong and Youssefi, 1976, Chakravarthy and Minker, 1986, Yannakakis, 1981, Bernstein and Goodman, 1981) The other is based on cost estimation of different strategies as in (Astrahan et al. 1976, Griffiths et al. 1979, Chamberlin et al. 1981, Daniels et al. 1982, Mackert and Lohman, 1986, Whang and Krishnamurthy, ....

Wong, E. & Youssefi, K. (1976). Decomposition -- A Strategy for Query Processing. ACM Trans. on Database Systems 1, 3, 223--241, 1976.

....(not shown) we have compared reoptimization using saved state without usage pointers and the resulting performance was worse than replanning from scratch. 9 Related work The INGRES query optimization algorithm originally interleaved steps of constructing a query execution plan and executing it [45]. However, their approach was largely eclipsed by less flexible System R style optimizers. Only recently have Kabra and DeWitt demonstrated the utility of runtime reoptimization for standard database queries using a System R style optimizer [26] The Tukwila rule mechanism enables reoptimization ....

E. Wong and K. Youssefi. Decomposition: A strategy for query processing. ACM Transactions on Database Systems, 1(3):223, 1976.

....polynomial time unless P5NP, because checking whether the join of a set of relations is nonempty is an NP complete problem [Chandra and Merlin 1977] Thus, in general, we cannot expect to find an efficient algorithm for computing the join of some relations. Query optimizers [Selinger et al. 1989; Wong and Youseffi 1976] search the space of orderings of joins and try to select a cheap ordering that does not generate huge intermediate relations. Since the number of possible orderings could be very large, query optimizers reduce the search space by using various heuristics. One heuristic commonly used in the ....

....(R 1 3 R 2 ) 3 . 3 R n . This heuristic is of practical interest, because it allows us to keep only one temporary relation at any time. Another heuristic is to avoid Cartesian products because they tend to be expensive. Query optimizers in many well known systems, such as INGRES [Wong and Youseffi 1976] and System R [Selinger et al. 1979] use one or both heuristics. Smith and Genesereth [1985] considered linear orderings of joins (conjunctions, to them) and gave an adjacency restriction rule that improves the cost of a join by locally swapping two adjacent relations. Swami [1989] and Swami ....

WONG, E., AND YOUSEFFI, K. 1976. Decomposition--A strategy for query processing. ACM Trans. Database Syst. 1, 3, (Sept.), 223--241.

....varying working principle, which are subject to a quantitative analysis in Section 5. Section 6 concludes the paper. 2 Problem Description The problem of determining good evaluation strategies for join expressions has been addressed from the development of the first relational database systems [WY76, YW79, SAC 79] The work in this area can be divided into two major streams: First, the development of efficient algorithms for performing the join itself, and second, algorithms that determine the nesting order in which the joins are to be performed. In this article, we shall be ....

....solutions are generally characterized by intermediate results with small cardinality. The minimum selectivity heuristic builds a left deep processing tree step by step while trying to keep intermediate relations as small as possible. In this regard, this resembles Ingres decomposition strategy [WY76] however, unlike the decomposion strategy, which considers only the operands cardinalities, the minimum selectivity heuristic makes use of the selectivity factor oe of the join R 1 Gamma R 2 to achieve small intermediate results. First, the set of relations to be joined is divided into two ....

E. Wong and K. Youssefi. Decomposition---A strategy for query processing. ACM Trans. on Database Systems, 1(3):223--241, 1976.

....of partitioning cost with little penalty. For example, once one arrives at the set of joins that minimize the number of repartitionings, one can order these joins based on other criteria, such as size. For a discussion of other such criteria, see [JK84, U82] for reviews and original papers [AHY83, BGWRR81, ES80, GS81, GS82, HY79, Schk82, WY76]. 2. Terminology A relation scheme R is a finite set of attributes A 1 , A n . Associated with each attribute A i is a domain denoted dom(A i ) A relation instance r on scheme R is a finite set of mappings t 1 , t m from R to the set of domains such that for each t r, t(A ....

E. Wong and K. Youssefi, "Decomposition -- a strategy for query processing" ACM TODS 1, 3 Sept. 1976, pp. 223-241.

....frequent joins, i.e. natural joins, are typically supported by (one or two) indices. In this case, Index Nested Loops is the strategy of choice as well. Our second result is to extend our approach to obtain a unified implementation for multiple joins in star queries of k#2 joins. As asserted in [WY76, KS86, Ull89, Cha91], a multiway join algorithm (like Ingres tuple substitution) is frequently better than any sequence of binary joins, especially when join attributes are indexed and each join increases result size. We show that a cascade of binary hierarchical joins (HJOIN) combines the advantages of two ....

....exploits the fact that the matching tuples found in relation R 2 are really determined only by the distinct join values. b) An early multiway Join Operator. There have been many suggestions to include multiway join within an optimizer s repertoire [KS86, Ull89, Cha91] The Wong Youssefi algorithm [WY76] implemented in the Ingres research prototype, cannot be simulated by a sequence of 2 way joins and is frequently more efficient than any sequence of 2 operand joins on flat tuples. Its behavior is: ####### # ### # # # ## ##### ## # ## # ## ### ## ## # ## # ######### # ######## # ## ....

Wong, E. and Youssefi, K., "Decomposition -- A Strategy for Query Processing", ACM TODS, vol. 1, no. 3, pp. 223--241, September

....(not shown) we compare re optimization using saved state without usage pointers and the resulting performance is worse than replanning from scratch. 7 Related work The INGRES query optimization algorithm originally interleaved steps of constructing a query execution plan and executing it [26]. However, their approach was largely eclipsed by less flexible System R style optimizers. Only recently have Kabra and DeWitt demonstrated the utility of runtime re optimization for conventional database queries using a System R style optimizer [15] The Tukwila rule mechanism enables ....

E. Wong and K. Youssefi. Decomposition: A strategy for query processing. ACM Transactions on Database Systems, 1(3):223, 1976.

....but that approach to query execution will create large temporary relations for the partial results. We will next discuss an approach to optimization to avoid this problem. 4. 3 Optimization Our approach to optimization is similar to the well known decomposition method of query processing [16, 22, 26]. An advantage of the decomposition approach is that during composition we do not have to use the original relations for the calculation of the partial queries. We can use the results of other partial queries instead. Because we join the partial results together, only the tuples of the partial ....

E.Wong and K.Youssefi. Decomposition --- A Strategy for Query Processing. ACM Trans. on Database Systems, Vol.1, No.3, p.223-241, 1976.

....varying working principle, which are subject to a quantitative analysis in Section 5. Section 6 concludes the paper. 2 Problem Description The problem of determining good evaluation strategies for join expressions has been addressed from the development of the first relational database systems [WY76, YW79, SAC 79] The work in this area can be divided into two major streams: First, the development of efficient algorithms for performing the join itself, and second, algorithms that determine the nesting order in which the joins are to be performed. In this article, we shall be ....

E. Wong and K. Youssefi. Decomposition---A strategy for query processing. ACM Trans. on Database Systems, 1(3):223--241, 1976.

.... Hong and Wong mention that decomposition at the early stage of MQP reduces the complexity of unnecessary combinations for queries with nothing in common (irrelevant queries) 7] Wong uses a matrix and connectivity algorithm to detect connected components for processing single queries in Ingres [13]. We have extended this matrix approach to connectivity detection for MQP. The problem of identifying common subexpressions is NP hard ( 10] 8] Therefore, Jarke indicates that multi relation subexpressions can only be addressed heuristically [8] Hall represents a query as a string of symbols ....

....nodes from distinct queries [11] Another issue in MQP is the representation and the processing of multiple queries. The multi graph is proposed for representing multiple SPJ type queries in [1] This multi graph can facilitate query processing by using Ingres instantiation and substitution ( 1] [13]) In [2] the multi graph was modified for representing the initial state of multiple queries. Our approach differs from these in many ways. First we have provided a simple yet effective technique to decompose a set of queries into unrelated sets [5] Once these sets are created, then the queries ....

Eugene Wong, and Karel Youssefi. "Decomposition - A Strategy for Query Processing," ACM Transactions on Database Systems, Vol. 1, No. 3, pp. 223-241, September 1976.

....optimizer as a result of the manner in which they integrate spatial data with nonspatial data. In particular, the underlying architecture may limit some feasible strategies for spatial query processing. GEOQL s spatial query optimizer [Ooi 1988] extends the well known query decomposition technique [Wong and Youssefi 1976] to handle spatial queries as well. However, GEOQL is biased towards the relational side. In particular, every relation is supported by at most one spatial attribute that is implicit and is always associated with the relation. Each query is decomposed into disjoint subqueries that consist entirely ....

Wong, E. and Youssefi, K. 1976. Decomposition - A strategy for query processing. ACM Transactions on Database Systems, 1, 3 (Sept.), 223--241.

....calculated, but that approach to query execution will create large temporary relations for the partial results. So Penguin has a query optimization algorithm to reduce the size of the temporary data. The algorithm is similar to the well known decomposition method of query processing [7] [13]. An advantage of the decomposition approach is that during composition we do not have to use the original relations for the calculation of the partial queries. We can use the results of other partial queries instead. Because we join the partial results together, only the tuples of the partial ....

E.Wong and K.Youssefi. "Decomposition --- A Strategy for Query Processing." ACM Trans. on Database Systems, Vol.1, No.3, pp. 223--241, 1976.

....technique to database querying. Some important steps were already been taken in this direction many years ago. The idea of a reducer is precisely to eliminate tuples not participating in the solution [BFMY83] This idea was applied by using semi joins in database querying, which dates back to [WY76, BC83] One motivation for this work was to minimise data transfer in distributed databases, but another motivation was to optimise query processing on a single (RAP) database machine. Indeed our approach addresses a problem identified in the conclusion of Bernstein and Chiu s paper [BC83] ....

E. Wong and K. Youseffi. Decomposition - a strategy for query processing. ACM Transactions on Database Systems, 1(3):223--241, 1976.

....2.2. 2 INGRES The INGRES relational database system and QUEL query language were developed, also in the mid70s, at the University of Berkeley [139] The query optimizer for the university version of INGRES was based on the decomposition of multi variable queries into single variable queries [152]. This is done interactively with query execution using two basic steps; detachment and tuple substitution. In the detachment step the query predicates are examined to find smaller queries involving one or two relations. These smaller queries are detached from the original query to form a series ....

....then, after manipulating the algebraic query in different ways, transforms that to an access plan in a language for object access. Almost all query optimizers (extensible or not) process a query through some fixed sequence of optimization stages (e.g. 45] 54] 56] 64] 129] 130] 142] [152]) An Epoq optimizer can simulate this kind of sequential processing in a straightforward way. For example, each stage of processing could be represented by a separate region, with a single level of control over those regions responsible for executing them in the desired order. The higher level ....

Eugene Wong and Karel Youssefi. Decomposition -- A Strategy for Query Processing. ACM Transactions on Database Systems, 1(3):223--241, September 1976.

....3. 5 Reducing phase( G ) G ; identity no optimization Sdd1( G ) SDD1 algorithm [Ber81b] Bc( G ) Bernstein Chiu algorithm [Ber81a] Yol( G ) Yu et al. algorithm [Yu84] Joining phase( G ) Sys R( G ) System R algorithm [Sel79] U ingres( G ) University Ingres algorithm [Won76] Exodus( G, RS ) Exodus rule optimizer [Gra87] RS is the rule set . From the above catalogs, it is easy to see how different implementations of components can arise. A set of potential implementations is defined by equation ( to be the cross product of the catalogs of algorithms that ....

E. Wong and K. Youseffi, `Decomposition - A Strategy for Query Processing', ACM Transactions on Database Systems, 1 #3 (Sept. 1976).

....use a binary table to find out what portions of the expression have to be computed to bring the materialized view up to date. To evaluate each SPJ expression associated with a row of the table, we can make use of some known algorithm such as QUEL s decomposition algorithm by Wong and Youssefi [WY76]. Once more, there is a possibility of saving computation by re using partial computations common to several rows in the table. We now present the outline of an algorithm to update SPJ views differentially. Algorithm 5.1 The input consists of: i) the SPJ view definition V = X (oe C (R 1 1 R 2 ....

Wong, Eugene, and Karel Youssefi, "Decomposition - A Strategy for Query Processing, " ACM Transactions on Database Systems, Vol. 1, No. 3, September 1976, pages 223--241.

....chooses a suboptimal plan, a database and query can be constructed to make that error look arbitrarily detrimental. Database queries are by definition 7 In fact, the pioneering designs in query optimization were more accurately described by their authors as schemes for query decomposition [Wong and Youssefi 1976] and access path selection [Selinger et al. 1979] 18 Delta J.M. Hellerstein ad hoc, which leaves us with a significant problem: how does one intelligently analyze the practical efficacy of an inherently rough technique over an infinite space of inputs Three approaches to this problem have ....

Wong, E. and Youssefi, K. 1976. Decomposition - A Strategy for Query Processing. ACM Transactions on Database Systems 1, 3 (September), 223--241.

....equivalence, that includes plans with different predicates, different group by and join predicates. In an attempt to find fast access paths for view processing, Roussopoulos [Rou82b] and [Rou82a] provides a framework for interquery analysis based on query graphs introduced by Wong and Youssefi ( WY76] Kim [Kim84] describes a two stage optimization procedure similar to that of [GM80] The unit of sharing in queries in Kim s proposal is a relation and does not extend to arbitrary expressions that are part of a query. Thus his work is applicable in a restricted setting of single relation ....

E. Wong and K. Youssefi. Decomposition: A Strategy for Query Processing. ACM TODS, 1(3):223--241, Sept 1976.

....DEPT objects. Taking into account additivity of the . operator with respect to its first argument, we can transform the query to a more optimal form (DEPT where ( P aris in LOC) EMPLOY S:EMP where JOB = clerk ) We can also show that the decomposition method, concerning conjunctive queries [WoYo76], can be generalized in our framework through the concept of the partial independence of subqueries. 7 Conclusion In this paper we have presented an approach to query languages based on a modification of the concepts known in programming languages. We believe, the approach makes possible to ....

E. Wong, K. Youssefi. Decomposition - A Strategy for Query Processing. ACM Transactions on Database Systems, Vol.1, No 3, 1976, pp.223-241

.... algorithms have been proposed for selection of a near optimal query plan based on a statistical description of the database and a cost model for query plan execution [Hall 1976, Jarke Koch 1984, Krishnamurthy et al. 1986, Selinger et al. 1979, Smith Chang 1975, Stonebraker et al. 1976, Wong Youssefi 1976, Yao 1979] One important aspect of local query optimization is the transformation of one query plan into an equivalent, but more efficient, query plan. The size of the search space of equivalent query plans for a snapshot query is determined in part by the algebraic equivalences available in the ....

Wong, E. and K. Youssefi. Decomposition - A Strategy for Query Processing. ACM Transactions on Database Systems, 1, No. 3, Sep. 1976, pp. 223--241.

.... any of the following algorithms: a) Dayal s algorithms and operators to translate an SQL query into an algebraic expression [8] b) Ganski Wong s recursive unnesting algorithm [10] to transform the SQL query to a conjunctive (SelectProject Join) query followed by the Wong Youssefi algorithm [23, 24] or System R techniques for access path selection [21] or (c) Ceri Gottlob s translation of SQL queries to relational algebra expressions [5] Dayal provides the most complete treatment for unnesting and optimizing SQL queries [8] If Dayal s algorithm, for translating a conventional SQL query to ....

Wong, E., and Youssefi, K. Decomposition - A strategy for query processing. ACM Transactions on Database Systems 1, 3 (September 1976), 223--241.

....to support the special case that all queries in a given column differ only by a collection of parameters have not yet been implemented. Although more sophisticated query processing algorithms have been constructed [SELI79, KOOI82] our implementation builds on the original INGRES strategy [WONG76]. The implementation of QUEL has been accomplished using this code because it is readily available for experimentation. Integration of our constructs into more advanced optimizers appears straightforward, and we discuss this point again at the end of this section. Figure 1 shows a diagram of the ....

....appears straightforward, and we discuss this point again at the end of this section. Figure 1 shows a diagram of the extended decomposition process. Detachment of onevariable queries that do not contain multiple dot or relation level operators can proceed as in the original INGRES algorithms [WONG76]. Similarly, the reduction module of decomposition is unaffected by our extensions to QUEL. In addition, tuple substitution is performed when all other processing steps fail. A glance at the left hand column of Figure 1 indicates that a test for zero variables must be inserted into the original ....

Wong, E., "Decomposition: A Strategy for Query Processing," ACMTODS, Sept. 1976.

....search the space of relevant access structures to select the best access structure for retrieving a relation. This approach was first described in System R[76] Ingres[80] uses a very unsophisticated plan generator which handles one variable queries only. The Ingres optimization algorithm[91] decomposes multi variable queries into single variable queries. 2.2 Plan Generation in Object Oriented Systems Plan generators for object oriented databases have continued where relational systems left off. Lanzelotte and her colleagues[59] state, database query optimization is finding an ....

Eugene Wong and Karel Youssefi. Decomposition -- a strategy for query processing. ACM Transactions on Database Systems, 1(2):223--241, September 1976.

....to the correlation predicates. In the next section, we document semantic transformations of join queries in hierarchical and objectoriented database systems to show the wide applicability of semantic query optimization. 6 Exploiting uniqueness in nonrelational systems Several researchers [6, 7, 10, 11, 19, 23] have studied ways to rewrite nested queries as joins to avoid a nestedloops execution plan. When the query is converted to a join, the optimizer is free to choose the most efficient join strategy while maintainingthe semantics of the original query; the assumption is that a nested loops strategy ....

Eugene Wong and Karel Youssefi. Decomposition---A strategy for query processing. acmtods, 1(3):223--241, 1976.

....reduces the search space of plans significantly. In contrast to System R, both the university version and the commercial version of Ingres 4 R 2 R 1 R 1 R 3 R 4 R 3 R 4 R 2 Figure 1.1: Deep Tree Plan v.s. Bushy Tree Plan allow query plan trees of all shapes, i.e. bushy tree plans [59, 29]. Figure 1.1 shows an example of a deep tree plan and a bushy tree plan for the same four way join query. There are advantages and disadvantages for both deep and bushy tree plans. Apparently, query optimization becomes more difficult as the set of possible plans grows. The set of bushy tree plans ....

E. Wong and K. Youssefi. Decomposition - a strategy for query processing. ACM Transactions on Database Systems, 1(3):223--241, September 1976. 135

No context found.

E. Wong and K. Yousse . Decomposition-a strategy for query processing. ACM Trans. on Database Systems, 1(3):223-241, 1976.

No context found.

E. Wong and K. Youssefi. Decomposition - a strategy for query processing. ACM Trans. on Database Systems, 1(3):223--241, 1976.

No context found.

E. Wong and K. Youssefi, "Decomposition - A Strategy for Query Processing," ACM Transactions on Database Systems, Vol. 1(3), pp. 223-241, (Sept. 1976).

No context found.

E. Wong and K. Youssefi. Decomposition: A strategy for query processing. ACM Transactions on Database Systems, 1(3):223, 1976.

No context found.

E. Wong and K. Youssefi, "Decomposition ---a Strategy for Query Processing", ACM Transactions on Database Systems,1,3, September 1976.

No context found.

W. Wong and K. Youseffi. Decomposition - A Strategy for Query Processing. ACM Transactions on Database Systems, Vol. 1, No. 3, pp. 223--241.

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