W. Hasan and H. Pirahesh. Query rewrite optimization in starburst. Research Report RJ6367, IBM, 1988.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....size is proposed. The enhancements deal with some special cases where the insensitivity claim does not hold, and consist of essentially introducing choose plan operators [GW89] The Starburst project has also considered incorporating a second optimization phase that chooses plans at run time [HP88]. To the best of our knowledge, however, no technique has been developed to find those plans. Also, Cornell and Yu [CY89] use an integer programming model to optimize queries and their buffer allocations in a transaction environment. Even though their concern is different from ours, their ....

Hasan W, Pirahesh H (1988) Query rewrite optimization in starburst. Technical Report RJ 6367, IBM Almaden Research Center

....A query optimization system that o ers a xed search strategy is always a rulebased system. It often implements a rule rewriter to perform equivalent transformation on the query expressions. Representatives of this kind of system include: the System R style Optimizer [15] Starburst project [16] [31], the Exodus Optimizer Generator [7] and the Volcano Optimizer Generator [8] The system R style Optimizer designs sets of rules to translate a query into a physical plan. One set of them is to convert the query into an algebraic tree. Other sets are used to generate access paths, join orders, ....

Waqar Hasan and Hamid Pirahesh. Query Rewrite Optimization in Starburst. Research Report RJ 6367 (62349), IBM, 1988. 176

....There can be pipelining of results between parent nodes and child nodes in the operator graph (data driven execution) as well as concurrent execution of nodes in the same level of the graph. Furthermore, even concurrent computation within one operator is possible. These issues are discussed in [19]. The aspect of sharing within a result set has already been mentioned: There may be several references pointing to the same atom within a set of molecules or even within one molecule. Hence, we had to develop a representation of the resulting molecule set which does not contain multiple copies ....

W. Hasan and H. Pirahesh, Query Rewrite Optimization in Starburst, IBM Research Report RJ6367, 1988.

....SQL nested queries at the SQL level [19] The leading tread behind the proposed transformations is to convert nested queries into joins so that a standard optimizer can work efficiently. However, there exist some major differences between nested queries and joins that are, as stated in [12], the creation of duplicates and the way empty tables are handled. Indeed, these differences are at the bottom of most of the bugs subsequently detected in the original algorithms [10, 17, 6, 8, 11, 12] To solve them and among other techniques, 8, 11] introduced outer joins. While these ....

....efficiently. However, there exist some major differences between nested queries and joins that are, as stated in [12] the creation of duplicates and the way empty tables are handled. Indeed, these differences are at the bottom of most of the bugs subsequently detected in the original algorithms [10, 17, 6, 8, 11, 12]. To solve them and among other techniques, 8, 11] introduced outer joins. While these algebraic operators are a nice solution to some of the above problems, they raise new issues. If a sequence of join operations can easily be reordered, joins and outer joins do not commute that easily and one ....

W. Hasan and H. Pirahesh. Query rewrite optimization in starburst. Research Report RJ6367, IBM, 1988.

....condition is done only if level 1 matching is satisfied. After the matching is found to be valid, the transformation is applied to the OAPT in order to get its equivalent OAPT according to that specific rule. The application of rules by rule based optimizers such as the EXODUS [11] and Starburst [15] optimizers is done by a pattern matching engine that matches subexpressions of a query against algebraic rules. Additionally, the firing of rules is dependent on the satisfaction of the conditions that involve user defined functions such as res(F; v) A major difference between the rules defined ....

W. Hasan and H. Pirahesh. Query rewrite optimization in Starburst. Technical Report TR RJ 6367, IBM Alamden Research Center, August 1988.

....size is proposed. The enhancements deal with some special cases where the insensitivity claim does not hold, and consist of essentially introducing choose plan operators [GW89] The Starburst project has also considered incorporating a second optimization phase that chooses plans at run time [HP88]. To the best of our knowledge, however, no technique has been developed to find those plans. Also, Cornell and Yu [CY89] use an integer programming model to optimize queries in a transaction environment and their buffer allocations simultaneously. Nonetheless, as their concern is different from ....

W. Hasan and H. Pirahesh. Query rewrite optimization in starburst. Technical Report RJ 6367, IBM Almaden Research Center, 1988.

....Shaw identifies a limited number of logical transformations for an object algebra [SZ90] The ability to precisely define the nature and pre conditions of logical transformations has led to their application using rule based systems. This technique can be used both to ameliorate queries [GD87, HP88] and to translate them into executable access plans [Fre87, GD87, Loh88] Generating access plans requires that (1) a well defined set of primitive query operations be implemented, and that (2) techniques exist for enumerating and evaluating alternate sequences of operations. The relational ....

....manipulating the class inheritance graph and determining type inclusion relationships. Chapter 6 presents a suite of equivalence preserving transformation rules for object algebra expressions. The rules are intended to serve as the rule base for a query rewrite system similar to that of Starburst [HP88] Three types of transformation rules are presented: identities, conditional rules and semantic rules. Identities are always applicable while conditional rules must meet a set of well defined conditions to be eligible. Semantic rules additionally utilize the semantics of the object oriented data ....

[Article contains additional citation context not shown here]

W. Hasan and H. Pirahesh. Query Rewrite Optimization in Starburst. Technical Report TR RJ 6367, IBM Almaden Research Center, August 1988.

....optimizer that can be built using various sets of rules. One set of rules is used to convert the query into an algebraic tree. Other sets of rules are used to generate access paths, join orderings, and join methods in that order. The optimizer developed as a part of the Starburst project [LFL88, HP88] uses a two step process to optimize queries. The first phase uses a set of production rules to heuristically transform the query into an equivalent query that (hopefully) offers both faster execution than the old query and is better suited for cost based optimization. In the second phase, query ....

Waqar Hasan and Hamid Pirahesh. "Query Rewrite Optimization in Starburst". Research Report RJ 6367 (62349), IBM, 1988.

....size is proposed. The enhancements deal with some special cases where the insensitivity claim does not hold, and consist of essentially introducing choose plan operators [GW89] The Starburst project has also considered incorporating a second optimization phase that chooses plans at run time [HP88] To the best of our knowledge, however, no technique has been developed to find those plans. Also, Cornell and Yu [CY89] use an integer programming model to optimize queries in a transaction environment and their buffer allocations simultaneously. Nonetheless, as their concern is different from ....

W. Hasan and H. Pirahesh. Query rewrite optimization in starburst. Technical Report RJ 6367, IBM Almaden Research Center, 1988.

....less than the corresponding overhead of completely optimizing queries at start up time. Finally,dynamic plans are highly robust when actual values of run time parameter values differ from their values expected at compile time. The prime difference of our approach from other work in this area [Ant93a, Ant93b, CAB93, DMP93, HaP88, HoS93, MHW90, OHM92, SKP88] is the extension of compile time dynamic programming with costs that cannot be compared until run time. Much of the previous work has focused on developing heuristics applied at start up time; therefore, there has been no guarantee of plan optimality.The same is true for (compile time) parametric ....

W.Hasan and H. Pirahesh, "Query Rewrite Optimization in Starburst", Comp. Sci. Res. Rep.,San Jose, CA, August 1988.

....with unknown parameters, but does not include a complete search strategy to identify the dynamic plans and the places where the choose plan operator should 22 be placed. The Starburst project has also considered incorporating a second optimization phase that chooses plans at run time [23], but no technique has been developed to find those plans. Also, 13] uses an integer programming model to optimize queries in a transaction processing environment and their buffer allocations simultaneously. However, in the end, only one plan is produced per query, and that plan is susceptible to ....

Waqar Hasan and Hamid Pirahesh. Query Rewrite Optimization in Starburst. Research Report RJ 6367 , IBM Almaden Research Center, August 1988.

....optimizer that can be built using various sets of rules. One set of rules is used to convert the query into an algebraic tree. Other sets of rules are used to generate access paths, join orderings, and join methods in that order. The optimizer developed as a part of the Starburst project [LFL88, HP88] uses a two step process to optimize queries. The first phase uses a set of production rules to transform the query heuristically into an equivalent new query that (hopefully) offers both faster execution than the old query and is better suited for cost based optimization. In the second phase, ....

Waqar Hasan and Hamid Pirahesh. "Query Rewrite Optimization in Starburst". Research Report RJ 6367 (62349), IBM, 1988.

....not its performance. For example, POSTGRES s rewrite may be used to implement user defined semantics for update of views. In contrast, our emphasis is on transformation for the purpose of optimizing query execution. An earlier design of the Starburst Query Rewrite rule system is reported in [HP88] 1.4 Structure of the Paper Section 2 presents the abstract representation of queries used by the rewrite rules. The rules themselves are presented in Section 3. Section 4 describes the rule engine designed for Query Rewrite. Summary and conclusions appear in Section 5. 2 Starburst s Query ....

....rules respectively, guaranteeing that SELMERGE will eventually get to be executed. We have chosen a relatively simple query to measure the effect of the above rule in the performance environment explained in Section 2. In practice, queries are typically more complicated, and the 3 As noted in [HP88] the importance of triggering this rule is emphasized when we remember that early relational systems such as System R supported only mergable views. Query CPU Time Elapsed time Before Rewrite 20 min 34.51 sec 24 min 19.80 sec After Rewrite 0 min 1.10 sec 0 min 7.20 sec Table 3: Example 1, ....

Waqar Hasan and Hamid Pirahesh. Query Rewrite Optimization in Starburst. Research Report RJ 6367 , IBM Almaden Research Center, August 1988.

....C code that executes the transformation in simple cases. In this section we give a sketch of our implementation. EMS is a part of the query rewrite phase of the Starburst optimizer. Rewrites are done by a (production) rule based system that encodes each query transformation as a rewrite rule ( HP88] A forward chaining engine traverses the query graph depth first (normally) applying rewrite rules. EMS is applied to graph elements representing table expressions, and it is applied to one table expression at a time. Multiple firings of the EMS rule, as the graph is traversed, cumulatively ....

W. Hasan and H. Pirahesh. Query Rewrite Optimization in Starburst. IBM Research report, RJ 6367 (62349), August 1988.

No context found.

W. Hasan and H. Pirahesh. Query rewrite optimization in starburst. Research Report RJ6367, IBM, 1988.

No context found.

W. Hasan and H. Pirahesh. Query rewrite optimization in starburst. Research Report RJ6367, IBM, 1988.

No context found.

W. Hasan and H. Pirahesh. Query rewrite optimization in starburst. Research Report RJ6367, IBM, 1988.

No context found.

W. Hasan and H. Pirahesh. Query rewrite optimization in starburst. Research Report RJ6367, IBM, 1988.

No context found.

W. Hasan and H. Pirahesh. Query rewrite optimization in starburst. Research Report RJ6367, IBM, 1988.

No context found.

W. Hasan and H. Pirahesh. Query rewrite optimization in starburst. Research Report RJ6367, IBM, 1988.

No context found.

W. Hasan and H. Pirahesh. Query rewrite optimization in starburst. Research Report RJ6367, IBM, 1988.

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