L. Fegaras. An experimental optimizer for OQL. Technical Report TR-CSE-97-007, CSE, University of Texas at Arlington, 1997.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....the typability of an OQL query is to transform it into its internal representations, i.e. the calculus and the algebraic form. Fegaras and Maier present a standard of effectiveness for a database calculus, before proposing monoid calculus as an intermediate representation for OQL queries [Feg94, Feg97, Feg98, FM95, FM00] They prove that monoid comprehensions conform to this standard by being expressive enough to represent OQL s constructs, manipulable (w.r.t. type checking, rewriting et cetera) evaluable (giving opportunities for alternative execution plans) and uniform in handling ....

L. Fegaras. An experimental optimizer for OQL. Technical Report TR-CSE-97-007, University of Texas at Arlington, 1997.

....compilation of queries is discussed in detail in the rest of this section. Their execution is described in the section 5. The key design decisions that have shaped the parallel OQL query compiler for OQL are: 1. The parallel optimiser is an extension of an existing non parallel optimiser for OQL [11]. It uses a two phase approach, in which non parallel aspects of optimisation take place before parallel aspects, allowing reuse of well established techniques for non parallel op timisation. 2. The optimiser is modular, in that components have been designed with well defined interfaces, so that ....

L. Fegaras. An experimental optimizer for OQL. Technical Report TR-CSE-97-007, CSE, University of Texas at Arlington, 1997.

....query might involve computation of the predicate C1 :a1:a2 : a m = C2 :b1 :b2 : b n . This can be easily observed to be a join of the extents of the classes that form the path expression. So, one of the optimization schemes followed in lambda DB, is to convert the path expression into a join[11], and then select from the resultant join table. But in BODHI, a specialized index structure, Path Dictionary, is provided for this purpose, which avoids the costly joins altogether. A Path Dictionary can be built over most often queried paths in the application schema. But the identities of ....

L. Fegaras. An experimental optimizer for oql. Technical Report TR-CSE-97-007, University of Texas at Arlington, 1997.

....Translation into Query Internal Form (QIF) GRAIL queries are intrinsically nested structures. However, nested language structures generally imply some evaluation order, so we follow a number of earlier query planners in unnesting the source query prior to query optimisation (Paton 1990, Fegaras 1997). The QIF is a list of query components, each of which is a tuple (Base, Variable, Criteria, Cost, Cardinality) representing the evaluation of part of the query. Base is the base concept of the component, Variable is the name of the variable used to store values retrieved as a result of evaluation ....

Fegaras L. An experimental optimizer for OQL. Technical Report TR-CSE-97-007, CSE, University of Texas at Arlington, 1997.

....no need to recourse to base objects and extents for certain kinds of update. Our solution works for MVs that refer to any ODL definable type (excluding array and dictionary collections) and are definable using the reduce, join, get, nest and unnest bulk operators of the monoid algebra proposed in [5, 6, 7] (excluding self joins) In terms of surface forms, this is roughly equivalent to select from where OQL queries (excluding, e.g. aggregations, set operations and quantifier constructions) but over a richer object data model. It is valid for any update operation defined in the ODMG standard. In ....

....both in research and in industrial circles. The chosen implementation strategy is to adopt an algebraic approach to update propagation and to rely on kernel level extensions for event monitoring. The query processing components that are assumed in this paper are those implemented in OPTGEN [5, 7] and lambda DB [6] systems. 2.1 ODMG Databases The ODMG standard for object oriented data management specifies, among other things, an object model (OM) an object definition language (ODL) and an object query language (OQL) ODL is used to declare a schema, which defines the valid application ....

[Article contains additional citation context not shown here]

L. Fegaras. An Experimental Optimizer for OQL. Technical Report TR-CSE-97-007, University of Texas at Arlington, 1997.

....in the literature, and most of them are fairly complex [4] as they introduce special constructs and distinguished types of operators with different functionalities, for expressing parallelism inside queries. Our parallel algebra is an extension of the physical algebra proposed by Fegaras in [6] to include a parallelism related operator and a pointer based join operator. We chose to extend this algebra because it is based on the concept of Monoids [11] which allow uniform treatment for collections and scalars, and also because the use of this algebra allows us to build the parallel ....

....operator. We chose to extend this algebra because it is based on the concept of Monoids [11] which allow uniform treatment for collections and scalars, and also because the use of this algebra allows us to build the parallel optimizer over the conventional OQL [3] optimizer implemented by Fegaras [6, 7]. In other words, we are using Fegaras work on non parallel ODMG query optimization as a starting point, and extending it with parallel optimization and evaluation capabilities. The paper is organised as follows. Section 2 describes the query processing architecture of the Polar parallel ....

[Article contains additional citation context not shown here]

L. Fegaras. An experimental optimizer for OQL. Technical Report TR-CSE-97-007, CSE, University of Texas at Arlington, 1997.

.... object oriented queries, most based on extensions to the nested relational algebra (see, e.g. 34] for a discussion) The ability of OQL queries to represent sets, bags or lists adds additional problems and has led to the development of representations such as the monoid comprehension calculus [34,35] that provide a canonical representation for OQL queries and a uniform treatment of operations that apply to all three collection types. Internal plan representation While most work represents a query plan as a tree over some appropriate query algebra, with the flow of data represented by the ....

....operator) 34] Cost functions are generally used to make a final translation into a physical algebra representing the operations that can be performed by a database. A common technique for this process is by using rewrite rules on algebraic structures with extra cost information attached to nodes [35 37]. Various search strategies have been used, including special purpose algorithms and combinations of branch and bound search with random techniques such as random walk, iterative improvement and simulated annealing [38] SIMS also uses rewriting techniques to optimise its partially ordered query ....

L. Fegaras. An experimental optimizer for OQL. Technical Report TR-CSE-97-007, University of Texas at Arlington, 1997. http://lambda.uta.edu/oqlopt.ps.gz.

....complemented by the algebraic optimization stage. ffl Where DBMS source code is available, the integrated architecture could benefit from an existing optmizer evaluator implementation for OQL. The object algebra underlying the DOQL translation approach is based on the algebras presented in [2, 8, 11]. The bulk operators of the algebra reflect the extent oriented nature of access to ODMG databases and also have a straightforward representation in terms of the algebraic interfaces supported by query optimization tools. The following bulk operators are supported: join(A,B,pred,outer) the join ....

....in the query processing flow. The choice of a combined approach (logical algebraic) to the translation of DOQL allows the exploitation of deductive optimization techniques and objectoriented optimization techniques, as well as laying the foundation for the use of query optimization tools like [11] to optimize the algebraic expressions. Acknowledgements: The first author is sponsored by Conselho Nacional de Desenvolvimento Cient ifico e Tecnol ogico CNPq (Brazil) Grant 200372 96 3. ....

Leonidas Fegaras. An experimental optimizer for OQL. Technical Report TR-CSE-97-007, CSE, University of Texas at Arlington, 1997.

....produced when translating queries without unnesting (since the predicates of the outer queries cannot be promoted into the inner queries) 8 Implementation We have already built a prototype OQL optimizer based on the framework described in this paper. It is described in detail elsewhere [Feg97] Our OQL optimizer is expressed in a very powerful optimizer specification language, called OPTL, and is implemented in a flexible optimization framework, called OPTGEN, which extends our earlier work on optimizer generators [FMS93] OPTL is a language for specifying query optimizers that ....

L. Fegaras. An Experimental Optimizer for OQL. University of Texas at Arlington Technical Report TR-CSE-97-007. Available at http://www-cse.uta.edu/¸fegaras/oqlopt.ps.gz, May 1997.

.... by using outer joins combined with grouping [13, 6] In most OODB algebras, where grouping is an explicitly supported operator and can appear in any place in an algebraic form, any form of query nesting can be unnested by promoting the inner queries one level up in the query nesting hierarchy [3, 4]. This results to a large flat join, which is the concatenation of all the joins from every level in the nested query. Furthermore, all unnesting methods proposed in the literature introduce additional operators, such as outer joins and grouping, thus increasing the number of operators to be ....

....random relational query graphs. Our preliminary results show that our method is clearly superior to the iterative improvement method. As a future work, we are planning to incorporate this algorithm to the experimental query optimizer for OQL being developed at the University of Texas at Arlington [4]. We are also planning to combine this algorithm with a local search technique, which improves small fragments of the query plan by performing a real cost analysis. Acknowledgements: This work is supported in part by the National Science Foundation under grant IRI 9509955. ....

L. Fegaras. An Experimental Optimizer for OQL. University of Texas at Arlington Technical Report TR-CSE-97-007. Available at http://www-cse.uta.edu/ ¸fegaras/oqlopt.ps.gz, May 1997.

.... avoided by using outer joins combined with grouping [12] In most OODB algebras, where grouping is an explicitly supported operator and can appear in any place in an algebraic form, any form of query nesting can be unnested by promoting the inner queries one level up in the query nesting hierarchy [3, 4]. This results to a large flat join, which is the concatenation of all the joins in every level in the nested query. Furthermore, all unnesting methods proposed in the literature introduce additional operators, such as outer joins and grouping, thus increasing the number of operators to be ....

....random relational query graphs. Our preliminary results show that our method is clearly superior to the iterative improvement method. As a future work, we are planning to incorporate this algorithm to the experimental query optimizer for OQL being developed at the University of Texas at Arlington [4]. We are also planning to combine this algorithm with a local search technique, which improves small fragments of the query plan by performing a real cost analysis. Acknowledgements: The author is grateful to Yannis Tollis for pointers into the literature on min cut algorithms and to Ramez ....

L. Fegaras. An Experimental Optimizer for OQL. University of Texas at Arlington Technical Report TR-CSE-97-007. Available at http://www-cse.uta.edu/¸fegaras/oqlopt.ps.gz., May 1997.

....current OODB languages and is easy to normalize, while the algebra is lower level and can be directly translated into the execution algorithms supported by database systems. Finally, we report on an implementation of our ideas using a powerful optimizer specification framework, called OPTGEN [10]. Since the merit of query unnesting cannot be judged in isolation from other optimizations, we combined unnesting with other optimization techniques, such as materialization of path expressions into joins [1] performing selections as early as possible, rearranging join orders, choosing access ....

....of the following rule: Gamma f(b) a (g(a) a:M=b:M g(b) Gamma f(a) a:M (g(a) where a=b are range variables in g(a) g(b) 6 Building the Optimizer We have already built a prototype OQL optimizer based on the unnesting algorithm described in this paper. It is described in detail in [10]. Our OQL optimizer is expressed in a very powerful optimizer specification language, called OPTL, and is implemented in a flexible optimization framework, called OPTGEN, which extends our earlier work on optimizer generators [13] OPTL is a language for specifying query optimizers that captures a ....

L. Fegaras. An Experimental Optimizer for OQL. University of Texas at Arlington Technical Report TRCSE -97-007. Available at http://www-cse.uta.edu/ ¸fegaras/oqlopt.ps.gz, May 1997.

....languages. Our calculus, called the monoid comprehension calculus, has already been shown to capture most features of ODMG OQL [Cat94] and is a good basis for expressing various optimization algorithms concisely, including query unnesting [Feg98] and translation of path expressions into joins [Feg97] In this paper, we extend our framework to handle object identity and object updates. 1.1 Object Identity Complicates Query Optimization Object oriented programming is based on side effects, that is, on the modification of the object store. Even though modern OODBs provide declarative query ....

L. Fegaras. An Experimental Optimizer for OQL. University of Texas at Arlington Technical Report TR-CSE-97-007. Available at http://www-cse.uta.edu/¸fegaras/oqlopt.ps.gz, May 1997.

No context found.

L. Fegaras. An experimental optimizer for OQL. Technical Report TR-CSE-97-007, CSE, University of Texas at Arlington, 1997.

No context found.

L. Fegaras. An experimental optimizer for OQL. Technical Report TR-CSE-97-007, University of Texas at Arlington, 1997. http://lambda.uta.edu/oqlopt.ps.gz.

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