C. Y. Chan and B. C. Ooi. Efficient Scheduling of Page Access in Index-Based Join Processing. IEEE Transactions on Knowledge and Data Engineering, 9(6):1005--1011, November/December 1997.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....solution exists for this problem. Solutions in the literature use a clustering method that groups pages in one or both tables involved in the join to reduce total page accesses. Available heuristics either group the pages of a single table via sorting [41] or use incremental clustering methods [7, 11, 33]. Our Contribution: We introduce two new heuristics for this problem. One heuristic uses the clustering method to group the pages in one table, generalizing the sorting based heuristic for joins. The other heuristic uses clustering for the pages of both tables. The former generalizes and ....

....and polygon data sets of California records as in Figure 6. Readers may note our use of real spatial data sets from the Sequoia 2000 [40] benchmark due to the interest in spatial join indices in contrast to the use of synthetic data sets, such as randomly generated graphs, in much related work [11, 7, 33]. We plan to use additional data sets, both real and synthetic, to expand the experimental evaluation in future work. Now, consider the following queries: Q.A. For each place in the Point table, find N nearest croplands from the Polygon table . Q.B. For each place in the Point table, find all ....

[Article contains additional citation context not shown here]

C. Y. Chan and B. C. Ooi. Efficient Scheduling of Page Access in Index-Based Join Processing. IEEE Transactions on Knowledge and Data Engineering, 9(6):1005--1011, November/December 1997.

....spans more than one cell. It also incurs an additional cost for reordering the whole dataset according to the epsilon grid order. Since sequence datasets can not be reordered, the efficiency of this technique degrades for such datasets. Two of the earlier techniques closer to ours in spirit are [11] and [42] Chan and Ooi [11] model the page access sequence as a concatenation of segments. Each segment is a sequence of pages that does not contain pages that do not contribute to the join. The experimental results show that the number of page accesses of this technique is only 10 less than the ....

....It also incurs an additional cost for reordering the whole dataset according to the epsilon grid order. Since sequence datasets can not be reordered, the efficiency of this technique degrades for such datasets. Two of the earlier techniques closer to ours in spirit are [11] and [42] Chan and Ooi [11] model the page access sequence as a concatenation of segments. Each segment is a sequence of pages that does not contain pages that do not contribute to the join. The experimental results show that the number of page accesses of this technique is only 10 less than the naive greedy technique. ....

C.C. Chan and B.B. Ooi. Efficient scheduling of page access in indexbased join processing. TKDE, 9(6):1005 1011, 1997.

....Section 9 presents the experimental results. We end with a brief discussion in Section 10. 2 Related work Joining two datasets is a costly operation. Current techniques reduce this cost by pruning pairs of data Without Index With Index point data [6, 7, 19, 44] 8, 11, 24] spatial data [3, 12, 29, 30, 36, 38, 45] [5, 20, 23, 31, 32] Table 1. A classification of join techniques. points that do not appear in the final join. They can be classified into two groups based on the data structures they use: 1) no index is built on the datasets, or 2) index is built only on at least one of the datasets. Another ....

....spans more than one cell. It also incurs an additional cost for reordering the whole dataset according to epsilon grid order. Since sequence datasets can not be reordered, the efficiency of this technique degrades for such datasets. Two of the earlier techniques closer to ours in spirit are [12] and [45] Chan and Ooi [12] modeled page access sequence as a concatenation of segments. Each segment is a sequence of pages that does not contain a premature page, where premature page is defined as a page that do not contribute to the join. The experimental results show that the number of page ....

[Article contains additional citation context not shown here]

C.C. Chan and B.B. OOi. Efficient scheduling of page access in index-based join processing. TKDE, 9(6):1005-- 1011, 1997.

....what page access sequence will require the minimum number of buffer pages 2. Given a fixed buffer size, what page access sequence will require the minimum number of page reaccesses The above problems are referred to as OPAS1 and OPAS2 respectively, for optimal page access sequence problems [6]. Both problems are believed to be NP Complete. All previous works have concentrated on finding a good solution to OPAS1, and then adapting it to OPAS2 by including a page replacement strategy when the maximum buffer size is reached. In this study, we present a new heuristic for OPAS2 that is not ....

....(OI I) 5] finds at each step the smallest number of fetches which would remove one page from the memory. The victim page when the buffer is full is the page with the smallest non resident degree that is not adjacent to the page being brought into the buffer. Chart and Ooi s Heuristic (COI I) [6] does not restrict itself to removing pages from the buffer. At each step, it looks for the smallest number of pages to be read in order to release any page, or the smallest minimal segment, and puts it into the buffer in order of descending resident degree. When the buffer is full, the page ....

B.C. OoI AND C. Y. CHAN, Efficient scheduling of page access in index-based join processing, IEEE Trans. Knowledge and Data Eng., 9 (1997), pp. 1005-1011.

....OPAS FB problem included AC, SC, FP, OM, CO, and Sorting. However, we did not include the Sorting method since it performed worse than AC on I O cost measures in our previous experiments. Editor: The latest paper on the topic is Efficient Scheduling of Page Access in Index Based Join Processing [3]. It was published recently and did not have any newer citation. As far as we know, the citation list is up to date and complete. ffl Comment: Performance experiments using only one dataset are not very convincing. Then, the reader is not able to decide whether the presented algorithms are ....

....missing Response: We added the following text at Section 2.3. 1: Readers may note the use of real spatial data sets from Sequoia 2000 [8] benchmark due to the interest in spatial join indices in contrast to the use of synthetic data sets, e.g. randomly generated graphs in much of related work [5, 3, 7]. We plan to use additional data sets, both real and synthetic, to expand the experimental evaluation in future work. We have added the following text in section 2.3.2: Note that in Figure 8(b) the increase of buffer size from 15 to 20 of the smaller relation did not reduce the number of ....

[Article contains additional citation context not shown here]

C. Y. Chan and B. C. Ooi. Efficient Scheduling of Page Access in Index-Based Join Processing. IEEE Transactions on Knowledge and Data Engineering, 9(6):1005--1011, November/December 1997.

....time solution exists for this problem. Solutions in literature use a global clustering method to group pages in one or both tables involved in join to reduce total page access. Available heuristics either group pages of a single table via global sorting [33] or use incremental clustering methods [6, 9, 27]. We introduce two new heuristics for this problem. One heuristic uses global clustering method to group pages in both tables. The other one uses global clustering for pages of a single table using join index information. Both methods use min cut graph partitioning y as clustering algorithm. The ....

....of edges connecting p to a node not in the buffer is minimal. If a node in the buffer has to be replaced to make room for the new node, then choose the node that (a) is connected to the fewest number of nodes outside the buffer, and (b) is not connected to the new node. Symmetric Heuristic: Chan [6] first, a SelectSegment heuristic selects the minimal segment that has the shortest non resident length. From this minimal segment, a SelectPage heuristic chooses 5 the page that has the largest resident degree from this segment. For the selection of victim pages for replacement when buffer is ....

[Article contains additional citation context not shown here]

Chee Yong Chan and Beng Chin Ooi. Efficient Scheduling of Page Access in Index-Based Join Processing. IEEE Transactions on Knowledge and Data Engineering, November/December 1997.

....solution exists for this problem. Solutions in the literature use a clustering method that groups pages in one or both tables involved in the join to reduce total page accesses. Available heuristics either group the pages of a single table via sorting [34] or use incremental clustering methods [5, 7, 28]. Our Contribution: We introduce two new heuristics for this problem. One heuristic uses the clustering method to group the pages in one table, generalizing the sorting based heuristic for joins. The other heuristic uses clustering for the pages of both tables. The former generalizes and ....

....table, find all croplands which are within a distance D . The spatial join of these two queries produces sets of join indices and such join indices are of interest in spatial data mining for neighborhood indexing [6] The value of N and D can be increased decreased for adjusting the edge ratio [5]. Give a join graph G = V R ; VS ; E) the edge ratio of G, denoted by Theta, is defined as the ratio of the total number of edges in G to the maximum possible number of edges in G if it is a fully connected graph; i.e. Theta = jEj jVR jjV S j . The edge ratio provides a measure of the ....

[Article contains additional citation context not shown here]

Chee Yong Chan and Beng Chin Ooi. Efficient Scheduling of Page Access in Index-Based Join Processing. IEEE Transactions on Knowledge and Data Engineering, November/December 1997.

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