KR102005343B1 - Partitioned space based spatial data object query processing apparatus and method, storage media storing the same - Google Patents

Abstract

The present invention relates to a partition space-based spatial data object query processing apparatus and method, comprising: a partition space generation unit for generating a plurality of partition spaces by partitioning a data space including at least one spatial data object and the plurality of partition spaces And an index tree generation unit for generating an index tree based on minimum boundary rectangle (MBR) information including all of the information about the spatial data object included in the partition and the partition. Therefore, the present invention can efficiently partition the data space and perform efficient query processing using an appropriate table schema.

Description

Spatial data object query processing apparatus and method based on partition space, and recording medium recording the same {PARTITIONED SPACE BASED SPATIAL DATA OBJECT QUERY PROCESSING APPARATUS AND METHOD, STORAGE MEDIA STORING THE SAME}

The present invention relates to a spatial data object query processing technology, and more particularly, to a partition space-based spatial data object query processing apparatus and method capable of efficiently partitioning a data space and performing an efficient query processing using an appropriate table schema. It is about.

The key to effective similarity query processing on multidimensional data such as spatial data is to store spatially adjacent objects in a physically close location and access only the data belonging to the area required for retrieval. However, cloud computing-based frameworks commonly used to deal with big data, such as Hadoop, do not take into account the spatial proximity between objects, so they do not contain data with a large amount of false positives when processing queries. Search.

Spatial data object queries can include range queries and k-Nearest neighbor queries, and indexing methods using NoSQL database management systems such as HBase provide fast data random access and efficient data in distributed file systems. It can provide an effective framework for updating. The range query takes as input a query point q and a query radius r and returns as a result a set of data objects whose distance from the query point q is less than r. The kNN query takes as input a query point q and the number of nearest neighbors k and returns as a result the set of k data objects closest to the query point q.

The problem with indexing in HBase is that it is designed to support only one-dimensional row keys. Therefore, linearization techniques are commonly used to store spatial data in HBase systems. This method subdivides the data space into grid-shaped cells and uses the z-order to sort the cells sequentially. The row key of the spatial data can be generated using the z-order number of the corresponding cell.

4 is an exemplary diagram illustrating a process of processing a range query using a linearization technique. Referring to FIG. 4, when a range query 410 is executed, minimum and maximum z-order values within a query range may be calculated. Next, the query can be processed by retrieving rows whose row key is within the z-order range. However, the result of the linearization technique can often include a large number of false positives 430 because spatial proximity cannot be fully guaranteed.

5 is an exemplary diagram illustrating a process of processing a range query using a multidimensional indexing layer. Referring to FIG. 5, when the range query 510 is executed, a false positive may be reduced by first searching an index layer 530 and accessing only rows having a row key associated with a related area. However, this method may still require access to a large amount of false positives in query processing because of the large unit of spatial partitioning.

Korean Patent Laid-Open No. 10-2016-0004781 (2016.01.13) relates to an ontology query processing method using a big data framework, and stores a class type as a prefix in each node stored in a table. Classify and store the schema to define each table column family and properties, and analyze the query based on the improved table schema of the HBase-based RDF repository to apply the HBase filter to search. It is possible to process queries efficiently by reducing the number of blocks, and to process nested triple pattern queries that cannot be processed by the existing query processor.

Korean Patent No. 10-1117709 (2012.02.10) relates to a multi-dimensional histogram method using a minimum data-uneven cover of a spatial partition tree and a recording medium storing a program for executing the same. Based on the inequality of the data objects in the partitioned space formed by dividing by, the minimum data-unbalance cover is determined, and a bucket of the histogram is created based on this, so that the data objects are equal, unlike the conventional multidimensional histogram method. Even in a non-distributed situation, it is effective to secure the accuracy of the estimation calculation for the selectivity of the region query.

Korean Patent Publication No. 10-2016-0004781 (2016.01.13) Korea Patent Registration No. 10-1117709 (2012.02.10)

An embodiment of the present invention is to provide a partition space-based spatial data object query processing apparatus and method capable of efficiently partitioning a data space and efficient query processing using an appropriate table schema.

An embodiment of the present invention provides an apparatus and method for processing a spatial data object based on a partition space capable of generating an index tree based on a plurality of partition spaces divided by a recursive quadrant and a minimum bounding rectangle of the spatial data object. To provide.

An embodiment of the present invention is to provide an apparatus and method for processing a spatial data object based on partition space that can effectively reduce false positives by processing similar queries using an efficient index tree.

In one or more embodiments, a partition space-based spatial data object query processing apparatus may include a partition space generation unit configured to generate a plurality of partition spaces by partitioning a data space including at least one spatial data object and information on the plurality of partition spaces. And an index tree generation unit for generating an index tree based on minimum boundary rectangle (MBR) information including all spatial data objects included in the partition.

The partition space generator may repartition the partition space recursively based on the center point of the partition space until the density of the corresponding spatial data object becomes less than or equal to a specific criterion for each of the plurality of partition spaces.

The index tree generation unit indexes the index based on a table including row data including a key value generated based on the information about the partition space and a value generated based on the minimum boundary rectangle information. You can create a tree.

Each time the partition space is generated, the index tree generator displays a partition space in a direction close to the origin as 0 and a partition space in a distant direction as 1 for each axis, and connects bits for each axis. You can generate key values.

When the recursive repartition of the partition space occurs, the index tree generation unit may display the partition information by connecting the previous partition information and the information on the current repartition space.

The index tree generating unit may generate the index tree by generating an index table for storing the row data of an internal node and a data table for storing the row data of a leaf node.

The spatial data object query processing apparatus based on the partition space may further include a query processing unit that processes a spatial data object query using the index tree.

The query processing unit may process a range query including a query point and a query radius or a kNN query including a query point and a nearest neighbor number using the index tree.

Among the embodiments, the spatial data object query processing method based on the partition space in the spatial data object query processing method performed in the spatial data object query processing apparatus based on (a) comprising at least one spatial data object. Generating a plurality of divided spaces by dividing a data space; and (b) minimum boundary rectangle (MBR) information including all information about the plurality of divided spaces and spatial data objects included in the divided spaces. Generating an index tree based on the;

The step (b) may be performed based on a table including row data including a key value generated based on the information about the partition space and a value value generated based on the minimum boundary rectangle information. It may be a step of generating an index tree.

In step (b), each time the partition space is generated, the partition space in the direction close to the origin is represented by 0 and the partition space in the distant direction is 1 for each axis, and the bits for each axis are connected. The key value may be generated.

In the step (b), when recursive repartitioning of the divided space occurs, the step of displaying the previous partitioning information and information on the current repartitioned space may be displayed in a manner of being connected.

The step (b) may be the step of generating the index tree by creating an index table for storing the row data for an internal node and a data table for storing the row data for a leaf node.

The partition space based spatial data object query processing method may further include (c) processing a spatial data object query using the index tree.

Step (c) may be a step of processing a range query including a query point and a query radius or a kNN query including a query point and a nearest neighbor number using the index tree.

In one or more embodiments, the recording medium may be a computer-executable recording medium for recording a spatial data object query processing method performed by a spatial data object query processing apparatus based on a partition space, and includes a data space including at least one spatial data object. The index tree is divided based on a process of generating a plurality of partition spaces by partitioning and information on minimum bounding rectangles (MBRs) including information on the plurality of partition spaces and spatial data objects included in the partition spaces. The process of creation.

The disclosed technique can have the following effects. However, since a specific embodiment does not mean to include all of the following effects or only the following effects, it should not be understood that the scope of the disclosed technology is limited by this.

An apparatus and method for processing a spatial data object based on a partition according to an embodiment of the present invention may generate an index tree based on a plurality of partition spaces and minimum bounding rectangles of spatial data objects partitioned through recursive partitioning. have.

The partition space-based spatial data object query processing apparatus and method according to an embodiment of the present invention can effectively reduce false positives by processing similar queries using an efficient index tree.

1 is a diagram illustrating a partition space based spatial data object query processing system according to an exemplary embodiment of the present invention.
FIG. 2 is a block diagram illustrating a spatial data object query processing apparatus in FIG. 1.
3 is a flowchart illustrating a process of processing a spatial data object query in the spatial data object query processing apparatus of FIG. 1.
4 is an exemplary diagram illustrating a process of processing a range query using a linearization technique.
5 is an exemplary diagram illustrating a process of processing a range query using a multidimensional indexing layer.
6 is an exemplary diagram illustrating a spatial partitioning process performed by a spatial data object query processing apparatus.
FIG. 7 is an exemplary diagram illustrating a structure of an index tree node generated by the index tree generator of FIG. 2.
8 is an exemplary diagram illustrating a table configuration of an index tree node generated by the index tree generator of FIG. 2.
9 is an exemplary view illustrating a range query processing process performed by a spatial data object query processing apparatus.

Description of the present invention is only an embodiment for structural or functional description, the scope of the present invention should not be construed as limited by the embodiments described in the text. That is, since the embodiments may be variously modified and may have various forms, the scope of the present invention should be understood to include equivalents capable of realizing the technical idea. In addition, the objects or effects presented in the present invention does not mean that a specific embodiment should include all or only such effects, the scope of the present invention should not be understood as being limited thereby.

On the other hand, the meaning of the terms described in the present application should be understood as follows.

Terms such as "first" and "second" are intended to distinguish one component from another component, and the scope of rights should not be limited by these terms. For example, the first component may be named a second component, and similarly, the second component may also be named a first component.

When a component is referred to as being "connected" to another component, it should be understood that there may be other components in between, although it may be directly connected to the other component. On the other hand, when a component is referred to as being "directly connected" to another component, it should be understood that there is no other component in between. On the other hand, other expressions describing the relationship between the components, such as "between" and "immediately between" or "neighboring to" and "directly neighboring to", should be interpreted as well.

Singular expressions should be understood to include plural expressions unless the context clearly indicates otherwise, and terms such as "comprise" or "have" refer to a feature, number, step, operation, component, part, or feature thereof. It is to be understood that the combination is intended to be present and does not exclude in advance the possibility of the presence or addition of one or more other features or numbers, steps, operations, components, parts or combinations thereof.

In each step, an identification code (e.g., a, b, c, etc.) is used for convenience of description, and the identification code does not describe the order of the steps, and each step clearly indicates a specific order in context. Unless stated otherwise, they may occur out of the order noted. That is, each step may occur in the same order as specified, may be performed substantially simultaneously, or may be performed in the reverse order.

The present invention can be embodied as computer readable code on a computer readable recording medium, and the computer readable recording medium includes all kinds of recording devices in which data can be read by a computer system. . Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, and the like. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.

All terms used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. Generally, the terms defined in the dictionary used are to be interpreted to coincide with the meanings in the context of the related art, and should not be interpreted as having ideal or excessively formal meanings unless clearly defined in the present application.

1 is a diagram illustrating a partition space based spatial data object query processing system according to an exemplary embodiment of the present invention.

Referring to FIG. 1, a partition-based spatial data object query processing system (hereinafter, referred to as a spatial data object query processing system) 100 includes a user terminal 110, a spatial data object query processing apparatus 130, and a database. 150 may be included.

The user terminal 110 may correspond to a computing device capable of inputting a spatial data object query and confirming a related response, and may be implemented as a smartphone, a notebook, or a computer, and is not limited thereto. It can also be implemented. The user terminal 110 may be connected to the spatial data object query processing apparatus 130 through a network, and the plurality of user terminals 110 including the user terminal 1 (110a) to the user terminal n (110c) may be a spatial data object. The query processing apparatus 130 may be connected at the same time.

The spatial data object query processing apparatus 130 may be implemented as a server corresponding to a computer or a program capable of receiving and processing a spatial data object query using an index tree generated by spatial partitioning. The spatial data object query processing apparatus 130 may be wirelessly connected to the user terminal 110 through Bluetooth, WiFi, or the like, and may exchange data with the user terminal 110 through a network.

The spatial data object query processing apparatus 130 may be implemented including a database 150 and may be implemented independently of the database 150. When implemented independently of the database 150, the spatial data object query processing apparatus 130 may be connected to the database 150 by wire or wirelessly to exchange data.

The database 150 is a storage device that can store various information necessary for processing spatial data object queries. The database 150 may store information about at least one range query or kNN query received from the user terminal 110 and a result obtained by processing the query, but is not limited thereto, and the spatial data object query processing apparatus ( 130 may store information collected or processed in various forms in the process of processing the spatial data object query.

In one embodiment, the database 150 may correspond to HBase capable of storing and managing spatial data. Here, HBase is Apache HBase, which may correspond to an open, non-relational database for the Hadoop platform. HBase has a data model similar to Google's Big Table, which provides fast random access to large amounts of structured data, and can provide real-time random read / write of data from the Hadoop Distributed File System (HDFS).

The database 150 may be composed of at least one independent sub-databases that store information belonging to a specific range, and may be configured as an integrated database in which at least one independent sub-databases are integrated into one. When composed of at least one independent sub-database, each sub-database may be wirelessly connected through Bluetooth, WiFi, and the like, and may exchange data with each other through a network. When the database 150 is configured as an integrated database, the database 150 may include a control unit for integrating respective sub-databases into one and managing data exchange and control flow between them.

FIG. 2 is a block diagram illustrating a spatial data object query processing apparatus in FIG. 1.

Referring to FIG. 2, the spatial data object query processing apparatus 130 may include a partition space generator 210, an index tree generator 230, a query processor 250, and a controller 270.

The partition space generator 210 may generate a plurality of partition spaces by partitioning a data space including at least one spatial data object. Here, the data space may correspond to the multidimensional space in which the spatial data object is defined. The spatial data object may be defined including attribute values corresponding to each dimension, and the partition space generator 210 may define a data space using a range of attribute values that the spatial data object may have.

For example, when the data space corresponds to a square space having a two-dimensional coordinate system and the length of each axis having a length of 100 from the origin, the partition space generating unit 210 has a position of (50, 50) which is the center point of the space. Based on this, it can be divided into four divided spaces based on a straight line parallel to each axis passing through the center point. Each divided partition may correspond to a square and may have the same size.

In one embodiment, the partition space generation unit 210 recursively divides the partition space with respect to each of the plurality of partition spaces based on the center point of the partition space until the density of the corresponding spatial data object is less than or equal to a specific criterion. Can be subdivided. Here, the density of the spatial data objects may correspond to the number of spatial data objects existing in a specific partition space. For example, when repartitioning until the number of spatial data objects existing in the partition is less than or equal to 4, the partition generation unit 210 corresponds to a specific partition if the partition includes 9 spatial data objects in the partition. The partition may be repartitioned into four partitions, and if there is a partition including four or more spatial data objects among the partitioned partitions, the partition may be repartitioned.

The index tree generation unit 230 may generate an index tree based on information about the partition space and minimum boundary rectangle (MBR) information including the partition data object. Here, the minimum bounding rectangle may correspond to the rectangle having the smallest width among the rectangles containing all the partition data objects. The index tree generator 230 may generate an index tree by using information about each of the minimum bounding rectangles including all the partitions and the spatial data objects existing in the partitions.

In one embodiment, the index tree generator 230 is composed of row data including a key value generated based on information on partition space and a value value generated based on minimum bounding rectangle information. You can create an index tree based on a table. For example, the index tree generation unit 230 may generate row data composed of (key, value) values and store the row data as one row of the HBase table.

In one embodiment, each time the partition space is generated, the index tree generator 230 displays the partition space in the direction close to the origin as 0 and the partition space in the far direction as 1 for each axis, and the bit for each axis. You can create a key value by concatenating (bit). For example, if there is a partition space generated through partitioning, the index tree generator 230 is '1' if the partition space is a direction far from the origin with respect to the x axis and a direction closer to the origin with respect to the y axis in the partition. , '0' and may be connected and finally displayed as '10'.

According to an embodiment, the index tree generator 230 may display the previous partition information and the information on the currently generated repartitioned space in a manner of concatenating repartitioned partitions. For example, when the index tree generator 230 is generated through partitioning and repartitioning occurs for the partition space indicated by '10', the four partitioned partitioned spaces are '00', '01', and '10', respectively. And '11', and may be connected to '10' which is the previous split information and finally displayed as '1000', '1001', '1010' and '1011'.

In an embodiment, the index tree generator 230 may generate an index tree by generating an index table that stores row data about an internal node and a data table that stores row data about a leaf node. The index table may store row data about an internal node, and the row data about an internal node may be a key value generated based on the information about the partition space generated by the partition generator 210, and the root of the corresponding internal node. The number of spatial data objects existing in the tree may be included, child node information of the corresponding internal node, and MBR information associated with each child node. The data table can store row data about leaf nodes, where the row data about leaf nodes is a key value generated based on information about the partition space associated with each leaf node, and a space within the partition space associated with that leaf node. May contain information about the data object.

The query processor 250 may process a spatial data object query using the index tree generated by the index tree generator 230. Here, the index tree may correspond to a quadrand-based MBR (Q-MBR) tree. The Q-MBR and the Q-BMR tree using the same will be described in more detail with reference to FIGS. 6 to 8.

In one embodiment, the query processor 250 uses the index tree generated by the index tree generator 230 to query a range query including a query point and a query radius or a kNN query including a query point and a nearest neighbor number. Can be processed. More specifically, the query processor 250 searches the Q-MBR tree according to Breadth First Search (BFS) while searching the index, and calculates a row key for loading the next iteration. Can be.

In one embodiment, the query processing unit 250 may use three sets of N, I, and D to search the index for range query processing. Here, N can store the loaded node for searching in the current iteration. I may correspond to a set of row keys to be loaded from the index table in the next iteration. D may correspond to a set of row keys to load from the data table.

The query processor 250 may insert the row key of the root node into I and then load the node from the index table to start the search. When the query processor 250 knows the maximum depth of the index tree, the query processor 250 may insert a plurality of row keys for lower level nodes into I, instead of inserting only a single row key for the root node.

When the current node is an internal node, the query processor 250 may calculate a minimum distance between the MBR of the child node and the query point q, and insert a row key of a child whose distance is within the query radius r into I. The query processing unit 250 may search all nodes in N and then prefetch additional nodes using the row key stored in I and store the result in N for the next iteration.

When the query processor 250 reaches the leaf node during the index search, the query processor 250 may store the row key of the leaf node in D, and load the relevant spatial data object at the end of the query processing. If there is no node in I, the query processor 250 may load the spatial data object stored in the leaf node in D, and may calculate a distance between the spatial data object and the query point q and respond to the range query.

The query processor 250 may insert only a spatial data object whose distance from the query point q is less than or equal to the query radius r to the result set. The query processor 250 may provide the user terminal 110 with information about the spatial data object belonging to the result set as a query processing result.

In one embodiment, the query processor 250 may maintain two priority queues, N and Q, for kNN query processing. Here, N may store nodes in ascending order of the minimum distance from the query point to MBR, and Q may store kNN candidate sets in descending order.

The query processor 250 may process the kNN query in two steps. In the first step, the query processing unit 250 may search the Q-MBR tree to find the row key of the leaf node using the approximation range r. Here, the approximate range may correspond to a minimum distance that guarantees a sufficient number of spatial data objects to find the desired number of neighbors.

The query processor 250 may sequentially check nodes stored in N to determine whether the minimum distance has a child node smaller than r. The query processor 250 may insert the row key of the child node into the list I when the minimum distance of the child node is smaller than r. The query processing unit 250 may set the range r to be very large at an initial stage, and may update from N to the maximum distance of the node when the number of objects of a node closer than the current node exceeds k.

When the current node is a leaf node, the query processor 250 may store the row key of the current node in the list L and load the same together. If there are no more nodes in N, the query processor 250 may load the node in which the row key is stored in I. The query processing unit 250 may terminate the first step when there is no node in N and there is no row key in I. The query processor 250 may identify the query result set by loading the spatial data object of the leaf node stored in L and evaluating the distance after the index search. The query processor 250 may store the evaluated spatial data object in Q, and may return a query result set when there are no more spatial data objects to be evaluated.

The controller 270 controls the overall operation of the spatial data object query processing apparatus 130 and manages a control flow or data flow between the partition space generator 210, the index tree generator 230, and the query processor 250. can do.

3 is a flowchart illustrating a process of processing a spatial data object query in the spatial data object query processing apparatus of FIG. 1.

Referring to FIG. 3, the spatial data object query processing apparatus 130 may generate a plurality of partition spaces by dividing a data space including at least one spatial data object through the partition space generation unit 210 (step). S310). In one embodiment, the partition space generation unit 210 recursively divides the partition space with respect to each of the plurality of partition spaces based on the center point of the partition space until the density of the corresponding spatial data object is less than or equal to a specific criterion. Can be subdivided.

The spatial data object query processing apparatus 130 may generate the index tree based on the minimum bounding rectangle information including all the information about the partition space and the spatial data objects included in the partition space through the index tree generator 230. It may be (step S330). In one embodiment, the index tree generator 230 is composed of row data including a key value generated based on information on partition space and a value value generated based on minimum bounding rectangle information. You can create an index tree based on a table.

In one embodiment, each time the partition space is generated, the index tree generator 230 displays the partition space in the direction close to the origin as 0 and the partition space in the far direction as 1 for each axis, and the bit for each axis. You can create a key value by concatenating (bit). According to an embodiment, the index tree generator 230 may display the previous partition information and the information on the currently generated repartitioned space in a manner of concatenating repartitioned partitions. In an embodiment, the index tree generator 230 may generate an index tree by generating an index table that stores row data about an internal node and a data table that stores row data about a leaf node.

The spatial data object query processing apparatus 130 may process the spatial data object query by using the index tree generated by the index tree generator 230 through the query processor 250 (step S350). In one embodiment, the query processing unit 250 may process a range query including the query point and the query radius or a kNN query including the query point and the nearest neighbor number using the index tree.

6 is an exemplary diagram illustrating a spatial partitioning process performed by a spatial data object query processing apparatus.

Referring to FIG. 6, the spatial data object query processing apparatus 130 is formed by two axes having a maximum size of 80 and vertically intersecting with respect to the origin 611 through the partition space generating unit 210. Data space 610 may be defined. The partition space generation unit 210 may generate a plurality of partition spaces by dividing the data space 610 based on the center points 40 and 40 of the space. The partition generation unit 210 may repartition the partitions recursively until the density of the corresponding spatial data object is less than or equal to a specific criterion. In FIG. 6, the density may correspond to four. That is, the partition generation unit 210 may repartition the space recursively until the number of spatial data objects existing in the single partition is less than or equal to four.

The spatial data object query processing apparatus 130 may generate a minimum boundary rectangle (MBR) 613 for the spatial data object existing in each quadrant after dividing the data space 610. The spatial data object query processing apparatus 130 may generate a Q-MBR (Quadrand-based MBR) 630 based on the information about the partition space 631 and the MBR 633, and may store the information in the HBase database. . The spatial data object query processing apparatus 130 may store the Q-MBR 630 in an HBase table and use it as a building block of a hierarchical index tree.

Information about the partition space 631 can be used as a row key to reduce the update cost. Each column value can be stored in a separate key-value pair format, and updating a row key can result in a large number of insertions in that key-value pair. Therefore, frequently updated MBR 633 information cannot be used as a row key. The worst case might be when you create a new row key for the grouped spatial data objects and regenerate all the key-value pairs when the update occurs. Thus, the MBR 633 information can be stored in a row that is relatively difficult to update and can be used for distance calculation during spatial query processing.

FIG. 7 is an exemplary diagram illustrating a structure of an index tree node generated by the index tree generator of FIG. 2.

Referring to FIG. 7, the spatial data object query processing apparatus 130 may generate an index tree for generating and managing a Q-MBR through the index tree generator 230. Here, the index tree may correspond to a Q-MBR tree and may be implemented as an HBase table or a memory index. The Q-MBR tree may be similar to a quad tree structure.

In an embodiment, the index tree generator 230 generates an index tree by generating an index table for storing row data about an internal node and a data table for storing row data about a leaf node. Can be generated. Internal nodes can be stored in an index table and consist of the quadrant information of the node, the MBRs of children (MBRs) of the child nodes, and the number of objects contained in the subtree. Can be. Leaf nodes may be stored in a data table and may comprise a quadrant, a number of objects of a leaf node, and a list of spatial data objects.

8 is an exemplary diagram illustrating a table configuration of an index tree node generated by the index tree generator of FIG. 2.

Referring to FIG. 8, quadrant information may be used to directly access the spatial data object stored in the data table instead of the pointer to the spatial data object existing in the leaf node region. The configuration of the table included in FIG. 8 includes an example of a hierarchical Q-MBR index structure for the spatial data object shown in FIG. Each Q-MBR may correspond to a leaf node of the index tree, and the Q-MBR of the parent node may be represented by a quadrant including the Q-MBRs of the child node and the MBR.

The '#objects' value of the 'Meta family' item in the index table indicates the number of spatial data objects in the tree whose node is the root node. For example, because the value of '#objects' is 15 for a node with a value of 'Rowkey' item 'Root', the number of spatial data objects in the tree is 15.

9 is an exemplary view illustrating a range query processing process performed by a spatial data object query processing apparatus.

Referring to FIG. 9, the spatial data object query processing apparatus 130 may process a range query through the query processing unit 250. The query processing unit 250 may insert the row key of R0 into L, and may start by loading it into N. Since the two child nodes R2 and R3 overlap each other in the query range, the query processor 250 may insert the row keys of R2 and R3 into I in the first iteration and load them together from the index table. Equally, the row keys of R9 and R6 can be inserted and loaded in the second iteration. Since R9 and R6 are leaf nodes, the query processing unit 250 may examine the spatial data objects of R9 and R6 for the response to the range query in the next iteration. The result set R includes two spatial data objects p1 and p2, and since there are no nodes to search, the query processor 250 may terminate the search.

Although described above with reference to a preferred embodiment of the present invention, those skilled in the art will be variously modified and changed within the scope of the invention without departing from the spirit and scope of the invention described in the claims below I can understand that you can.

100: Spatial Data Object Query Processing System based on Partition
110: user terminal 130: spatial data object query processing apparatus
150: database
210: partition space generation unit 230: index tree generation unit
250: query processing unit 270: control unit
410: range query 430: affirmative error
510: range query 530: index layer
610: data space 611: origin
613: Minimum bounding rectangle 630: Q-MBR
631: partition space 633: MBR

Claims (16)

A partition space generation unit configured to generate a plurality of partition spaces by partitioning a data space including at least one spatial data object; And
Whenever the division space is created, the division space in the direction close to the origin is 0 for each axis and the division space in the distant direction is represented by 1, and the key value is connected by connecting bits for each axis. And an index tree generator for generating an index tree based on minimum boundary rectangle (MBR) information including all the information about the plurality of partitions and the spatial data objects included in the partitions. Spatial data object query processing apparatus based on partition space.
The method of claim 1, wherein the partition space generating unit
Partition data based on partition space, characterized in that the partition space is recursively divided based on the center point of the partition space until the density of the corresponding spatial data object is less than a specific criterion for each of the plurality of partition spaces. Object Query Processing Unit.
The method of claim 1, wherein the index tree generation unit
Generating the index tree based on a table composed of row data including the key value generated based on the information about the partition space and a value value generated based on the minimum boundary rectangle information; Spatial data object query processing apparatus based on the partition space.
delete The method of claim 1, wherein the index tree generation unit
Partitioning-based spatial data object query processing apparatus characterized in that when the recursive re-partition for the partition space is generated in a manner of connecting the previous partition information and the information on the current re-partitioned space.
The method of claim 3, wherein the index tree generation unit
And said index tree is created by creating an index table for storing said row data for internal nodes and a data table for storing said row data for leaf nodes.
The method of claim 1,
Spatial data object query processing apparatus based on the partition space further comprises a query processing unit for processing a spatial data object query using the index tree.
The method of claim 7, wherein the query processing unit
A spatial data object query processing apparatus based on partition space, which processes a range query including a query point and a query radius or a kNN query including a query point and a nearest neighbor number using the index tree.
In the spatial data object query processing method performed in the spatial data object query processing apparatus based on partition space,
(a) dividing a data space including at least one spatial data object to generate a plurality of partition spaces; And
(b) Whenever the division space is created, the division space in the direction close to the origin is 0 for each axis and the division space in the distant direction is 1, and a bit is connected for each axis. Generating an index tree and generating an index tree based on minimum bounding rectangle (MBR) information including all information about the plurality of partitions and spatial data objects included in the partitions. Spatial data object query processing method based on partition.
The method of claim 9, wherein step (b)
Generating the index tree based on a table composed of row data including the key value generated based on the information about the partition space and a value value generated based on the minimum boundary rectangle information; Spatial data object query processing method based on partitioning, characterized in that the step.
delete The method of claim 9, wherein step (b)
When the recursive repartition for the partition space occurs, displaying the partition information based on the partition information, characterized in that the step of connecting the previous partition information and the information on the current repartitioned space.
The method of claim 10, wherein step (b)
Generating the index tree by generating an index table for storing the row data for an internal node and a data table for storing the row data for a leaf node; .
The method of claim 9,
and (c) processing the spatial data object query using the index tree.
The method of claim 14, wherein step (c)
And processing a range query including a query point and a query radius or a kNN query including a query point and a nearest neighbor number using the index tree.
A computer-readable recording medium having stored thereon code for executing a spatial data object query processing method performed in a partition-based spatial data object query processing apparatus,
Generating a plurality of partition spaces by dividing a data space including at least one spatial data object; And
Whenever the division space is created, the division space in the direction close to the origin is 0 for each axis and the division space in the distant direction is represented by 1, and the key value is connected by connecting bits for each axis. And generating an index tree based on minimum boundary rectangle (MBR) information including all the information about the plurality of partitions and the spatial data objects included in the partitions. .

Images