F. Raab, editor. TPC Benchmark(tm) D (Decision Support), Proposed Revision 1.0. Transaction Processing Performance Council, San Jose, 1995.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....The cost model is introduced in Section 4. We present the algorithms along with the analysis of their performance guarantees in the following section. Experimental results are stated in Section 6. Finally, we present our conclusions in Section 7. 2 Motivation We use an example taken from TPC D [TPCD], a decision support benchmark to motivate the index selection problem. The example illustrates the complexity of this problem, and the difficulty of doing the selection well in the two step selection process. For this example, we use only a subset of the dimensions that exist in the schema of the ....

F. Raab, editor. TPC Benchmark(tm) D (Decision Support), Proposed Revision 1.0. Transaction Processing Performance Council, San Jose, CA 95112, 4 April 1995.

....whose basic E 379 elements are facts, dimensions and hierarchies. In this section we briefly recall the E 387 features of the DFM useful within this paper, with reference to the LINEITEM E 353 example shown in Figure 2, which models one of the star schemes included within E 373 the TPC D [15]. E 100 Mkt Segment E 313 Region Nation E 300 Year E 341 Order Priority E 262 Container E 151 MFGR E 208 Quarter E 316 LINEITEM E 184 Part E 185 Price E 182 Qty E 182 Extended Price E 214 Discount E 182 Discount Price E 213 Sum Charge E 198 Tax E 182 Brand E 207 Region ....

....problems generated so far (except E 371 those such that all their subproblems have already been generated and processed) the E 387 one for which the cost of the relaxation is minimum. E 225 5 Experimental tests E 157 The tests we have carried out are based on the well known TPC D benchmark [15], E 380 which features two related fact schemes (LINEITEM and PARTSUPPLIER) The number of primary fact instances generated for the two schemes is 6.000.000 and E 386 800.000, respectively; the total amount of data is about 1 Gbyte. E 286 We have tested our approach with five workloads, the ....

Raab, F. (ed.): TPC Benchmark(tm) D (Decision Support). Proposed Revision 1.0.<E-383> Transaction Processing Performance Council, San Jose, CA 95112 (1995)

....of interest when an aggregation query references a grouped view (a view containing a group by) Experimental results show that the technique is very beneficial for queries in the TPC D benchmark. 1 Introduction Aggregation is widely used in decision support systems. All queries in the TPC D[Raa95] benchmark contain aggregation. Efficient processing of aggregation queries is essential for performance in decision support applications and large scale applications. Permission to copy without fee all or part of this material is granted provided that the copies are not made or distributed for ....

....eager lazy group by. Eager group by performs eager aggregation on all tables containing aggregation columns. Lazy group by is its reverse transformation. The following examples illustrate the basic idea of eager group by and lazy group by. The examples are based on a subset of the TPC D database[Raa95]. The tables are defined in Appendix A. Example 1 : Find the total loss of revenue on orders handled by each clerk due to parts being returned by customers. Output clerk and loss of revenue. SELECT OCLERK, SUM(LEXTENDEDPRICE (1 LDISCOUNT) FROM LINEITEM, ORDERS WHERE OORDERKEY = LORDERKEY AND ....

[Article contains additional citation context not shown here]

F. Raab, editor. TPC Benchmark(tm) D (Decision Support), Working Draft 9.1. Transaction Processing Performance Council, San Jose CA, 95112-6311, USA, February 1995.

.... disk cylinders, and progressing up through synthetic models of video ondemand applications and the data staging and checkpointing phases of a parallel scientific file system, up to disk I Olevel traces of a number of real life workloads including an unaudited run of the TPC C database benchmark [31], an HP UX timesharing system, and a number of IBM AS 400 production systems. This set included workloads with constant, Poisson, and bursty arrivals, open and closed queueing models, and a large range of spatial locality measures and request sizes. More details of the workloads and how we ....

F. Raab, editor. TPC benchmark C, Standard Specification Revision 3.0. Technical report. Transaction Processing Performance Council, 15 Feb. 1995.

....parameter for a test run. If the degree of concurrency is n, n sort jobs would be started initially and as soon as one finished another one would be started. To get some basis for deciding on a distribution of sort sizes, we analyzed the sorts generated when running the TPC D benchmark queries [Raa95] More specifically, we analyzed the execution plans used by DB2 6000 version 2 for each of the 17 queries on a TPC D database with 26 indexes. We found a total of 55 sorts with the size distribution shown in Table 3. Table 3: TPC D Sort Sizes, scale factor 1.0 Input size Average No of Frerange ....

F. Raab. TPC Benchmark(tm) D (Decision Support), Working Draft 9.1. Transaction Processing Performance Council, San Jose CA, 95112-6311, USA, February 1995.

....model is introduced in Section 4. We present the algorithms along with the analysis of their performance guarantees in the following section. Experimental results are stated in Section 6. Finally, we present our conclusions in Section 7. 2 Motivating Examples We use an example taken from TPC D [TPCD95], a decision support benchmark to motivate the indexselection problem. The example illustrates the complexity of this problem, and the difficulty of doing the selection well in the two step selection process. For this example, we use only a subset of the dimensions that exist in the schema of the ....

F. Raab, editor. TPC Benchmark(tm) D (Decision Support), Proposed Revision 1.0. Transaction Processing Performance Council, San Jose, CA 95112, 4 April 1995.

....question above we want the total sales of a given part p to a given customer c for ALL suppliers. The query is answered by looking up the value in cell (p; ALL; c) 2 We use the TPC D database of size 1GB as a running example throughout this paper. For more details on this benchmark refer to [TPCD]. We have only discussed the presentation of the data set as a multi dimensional data cube to the user. The following implementation alternatives are possible: 1. Physically materialize the whole data cube. This approach gives the best query response time. However, precomputing and storing every ....

F. Raab, editor. TPC Benchmark(tm) D (Decision Support), Proposed Revision 1.0. Transaction Processing Performance Council, San Jose, CA 95112, 4 April 1995.

No context found.

F. Raab, editor. TPC Benchmark(tm) D (Decision Support), Proposed Revision 1.0. Transaction Processing Performance Council, San Jose, 1995.

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