Search operations in databases require special support at the physical level. This is true for conventional databases as well as spatial databases, where typical search operations include the point query (find all objects that contain a given search point) and the region query (find all objects that overlap a given search region). More

Two new spatial join operations, distance join and distance semijoin, are introduced where the join output is ordered by the distance between the spatial attribute values of the joined tuples. Incremental algorithms are presented for computing these operations, which can be used in a pipelined fashion, thereby obviating the need to wait for their completion when only a few tuples are needed. The algorithms can be used with a large class of hierarchical spatial data structures and arbitrary spatial data types in any dimensions. In addition, any distance metric may be employed. A performance study using Rtrees shows that the incremental algorithms outperform non-incremental approaches by an order of magnitude if only a small part of the result is needed, while the penalty, if any, for the incremental processing is modest if the entire join result is required.

In many massive dataset applications the data must be stored in space and query efficient data structures on external storage devices. Often the data needs to be changed dynamically. In this chapter we discuss recent advances in the development of provably worst-case efficient external memory dynamic data structures. We also briefly discuss some of the most popular external data structures used in practice.

The goal of the VIMSYS project is to create an information management system that can retrieve images or parts of images in addition to textual data. The system allows users to incrementally formulate a query starting from semantic pictorial objects. 11 allows incompletely specified and similarity-based queries. This paper presents a new layered data model and the design of the query processing unit for this system. The data model is a combination of object-oriented and functional models, and permits multiple representations of image entities. The knowledge module guides the user to progrcssively refine similarity-based queries and to query by defining new semantic objects and composing its attributes. Design considerations for the access structure of the system are also discussed, alongwith some of the problems of creating efficient mechanisms of access. 2

Abstract: Recently there has been an increasing interest in supporting bulk operations on multidimensional index structures. Bulk loading refers to the process of creating an initial index structure for a presumably very large data set. In this paper, we present a generic algorithm for bulk loading which is applicable to a broad class of index structures. Our approach differs completely from previous ones for the following reasons. First, sorting multidimensional data according to a predefined global ordering is completely avoided. Instead, our approach is based on the standard routines for splitting and merging pages which are already fully implemented in the corresponding index structure. Second, in contrast to inserting records one by one, our approach is based on the idea of inserting multiple records simultaneously. As an example we demonstrate in this paper how to apply our technique to the R-tree family. For R-trees we show that the I/O performance of our generic algorithm meets the lower bound of external sorting. Empirical results demonstrate that performance improvements are also achieved in practice without sacrificing query performance. 1

In this paper we investigate automated methods for externalizing internal memory data structures. We consider a class of balanced trees that we call weight-balanced partitioning trees (or wp-trees) for indexing a set of points in R d . Well-known examples of wp-trees include kd- trees, BBD-trees, pseudo-quad-trees, and BAR-trees. Given an efficient external wp-tree construction algorithm, we present a general framework for automatically obtaining a dynamic external data structure. Using this framework together with a new general construction (bulk loading) technique of independent interest, we obtain data structures with guaranteed good update performance in terms of I/O transfers. Our approach gives considerably improved construction and update I/O bounds for e.g. external kd-trees and BBD-trees.

Wh1h th persistent data of many advanced database applications,such as OLAP and scientific studies, arech1EEV erized by very hry dimensionality, typical queries posed onthj# data appeal to a small number of relevant dimensions. Unfortunately, th multi-dimensional access meths1 designed forh1##FD/1T - sional data perform rathm poorly forth1S partially specified queries.Th retrievaltech ique proposed in th1 paper uses a combination of two complementary measures to support e#cient processing of partial queries over hEV/1TSjE sional data. First, an elaborate storage organization, calledth inverted space, allowsth system administrator to controlth size of individual indexes in order to avoidth negative impact of extremely hly data dimensionality on th retrieval performance. Second, a new indexing structure, whe, is designed to supportth inverted-space storage organization, enables e#cient query processing in projected spaces with moderate dimensionality. Th1 indexingmech&/ sm is a general-purpose point access meths th e#ectively attacksth limitations of KDB-trees in spaces with many dimensions, whns preserving th simplicity and relatively good performance of th later structure in low-dimensional spaces.Th analytical and experimental resultsshS thS th new indexingsch me outperforms twooth# variants of KDB-trees investigated in th paper for both fully and partially specified queries. # 2002 Elsevier Science B.V. Allrigh s reserved. Keywords: Database systems; Accessmeths1x Data dimensionality; Performance analysis www.elsevier.com/locate/datak Data & Knowledge Engineering 42 (2002) 1--21 * CorrespondingauthFD Tel.: +1-312-567-5343; fax: +1-312-567-5067.

Abstract. Emerging database applications require the use of new indexing structures beyond B-trees and R-trees. Examples are the k-D tree, the trie, the quadtree, and their variants. They are often proposed as supporting structures in data mining, GIS, and CAD/CAM applications. A common feature of all these indexes is that they recursively divide the space into partitions. A new extensible index structure, termed SP-GiST is presented that supports this class of data structures, mainly the class of space partitioning unbalanced trees. Simple method implementations are provided that demonstrate how SP-GiST can behave as a k-D tree, a trie, a quadtree, or any of their variants. Issues related to clustering tree nodes into pages as well as concurrency control for SP-GiST are addressed. A dynamic minimum-height clustering technique is applied to minimize disk accesses and to make using such trees in database systems possible and efficient. A prototype implementation of SP-GiST is presented as well as performance studies of the various SP-GiST’s tuning parameters. Keywords: space-partitioning trees, spatial databases, extensible index, generalized search trees, clustering

Abstract: Recently there has been an increasing interest in supporting bulk operations on multidimensional index structures. Bulk loading refers to the process of creating an initial index structure for a presumably very large data set. In this paper, we present a generic algorithm for bulk loading which is applicable to a broad class of index structures. Our approach differs completely from previous ones for the following reasons. First, sorting multidimensional data according to a predefined global ordering is completely avoided. Instead, our approach is based on the standard routines for splitting and merging pages which are already fully implemented in the corresponding index structure. Second, in contrast to inserting records one by one, our approach is based on the idea of inserting multiple records simultaneously. As an example we demonstrate in this paper how to apply our technique to the R-tree family. For R-trees we show that the I/O performance of our generic algorithm meets the lower bound of external sorting. Empirical results demonstrate that performance improvements are also achieved in practice without sacrificing query performance. 1

We present an order-preserving general purpose data structure for binary data, the LPC-trie. The structure