JP2013218490A - Information storage/search device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve the problem that since it is necessary to achieve search expression standby type search processing by storing and searching preliminarily registered search expressions by such a basic system as linear search in the case of achieving uTupleSpace, it is not possible to achieve any efficient device.SOLUTION: An information storage/search device configured to store and search information including at least dimensions and values includes: search expression storage means for storing applied search expressions including at least a plurality of dimensional value ranges in conditions by using a tree structure for search expressions capable of simultaneously performing range search on the basis of a plurality of dimensional values, and registering an entry having the range of values as the data entry of the tree structure; and a search expression search means for searching a registered entry group matching the search conditions from the tree structure for search expressions by using the data values of the applied information including at least the plurality of dimensional values or value ranges as the search conditions of the tree structure.

Description

  The present invention relates to a database construction method for handling sensor information.

  Conventionally, there has been uTupleSpace (Non-Patent Document 1 and Patent Document 1) as an information management system for handling information such as sensor measurement data, that is, sensor information. One of the features of uTupleSpace is that sensor information can be expressed more freely in the “Tuple format” in which “key = value” is arranged.

  Furthermore, another feature of uTupleSpace is that the measured data values are accumulated first and then the search formula is inserted to search for matching data values and return them to the search formula input source. In addition to data value storage type search processing as in the database management system of, it is possible to return the matching data value to the storage source of the search expression by storing the search expression first and registering the data value later It also supports search-type standby search processing, which can be handled without distinction. In terms of uTupleSpace, uTuple indicating a data value is referred to as EventActual (EA), and uTuple representing a search expression is referred to as EventFormal (EF).

  Conventionally, there has been a multi-dimensional search tree UBI-Tree (Patent Literature 2 and Patent Literature 3) as a data structure for collectively indexing data values having a plurality of values as in the u Tuple format.

  By using the UBI-Tree disclosed in these documents as a central processing method for accumulation and retrieval, data such as a normal database management system in which data values in the uTuple format are accumulated and retrieved according to the retrieval formula. The basic operation of the value accumulation type can be realized efficiently and easily.

JP 2012-003490 A JP 2011-170460 A JP 2011-170461 A

"Design and Implementation of New uTupleSpace Enabling Storage and Retrieval of Large Amount of Schema-less Sensor Data", Takayuki Nakamura, Keiichiro Kashiwagi, Yutaka Arakawa, Motonori Nakamura al., Second International Workshop on Enablers for Ubiquitous and Context-Aware Services on Sensor Networks (EUCASS 2011), July 2011

  Conventionally, efficient processing by UBI-Tree is possible for data value storage type search processing, but an efficient algorithm such as UBI-Tree is used for search type standby type search processing, which is a feature of uTupleSpace. There was a problem that it could not be used. This is because a tree structure algorithm such as UBI-Tree is generally devised and designed on the premise of storing data and extracting them with a search expression.

  For this reason, when realizing uTupleSpace, it is necessary to realize a search expression standby type search process by accumulating and searching pre-registered search expressions by an elementary method such as linear search, An efficient device could not be realized.

  The information accumulation and retrieval method of the present invention is an information accumulation and retrieval method for accumulating and retrieving information including at least a dimension and a value. A range of values can be simultaneously searched for by a plurality of dimensions, and a range of values is defined as a tree structure data entry Using a search expression tree structure capable of registering entries having a search expression storage procedure for storing a given search expression including at least a condition range of a plurality of dimensions as a condition, and at least a value or a value range of a plurality of dimensions A retrieval formula retrieval procedure for retrieving a registered entry group that matches the retrieval condition from the retrieval formula tree structure in order using the data value of the given information included as a retrieval condition for the tree structure.

In the information storage and retrieval method of the present invention, a retrieval request is received from the retrieval formula tree structure by receiving a registration request for the data value of the information and performing the retrieval formula retrieval procedure on the data value. A data value accumulation procedure for accumulating the data values in a data value tree structure capable of range search with a plurality of dimensional values, further comprising before or after the retrieval formula accumulation procedure,
In the retrieval formula accumulation procedure, a data value matching the retrieval formula requested for registration may be further retrieved from the data value tree structure.

  In the information storage and retrieval method of the present invention, for determining whether or not there is a match in the retrieval formula retrieval procedure, retrieval of the retrieval formula tree structure in which all the dimensions included in the registered entry to be judged are judgment targets is performed. All of the conditions are included, and at least the condition may be that the product set of the value range and the value range of both dimensions is not an empty set.

  In the information storage and retrieval method of the present invention, the retrieval formula tree structure may be a UBI-Tree algorithm.

  The information storage / retrieval apparatus of the present invention is an information storage / retrieval apparatus that accumulates and retrieves information including at least a dimension and a data value, and is capable of performing a range search based on values of a plurality of dimensions at the same time. Search expression storage means for storing a given search expression including at least a condition range of a plurality of dimensions using a search expression tree structure capable of registering entries having a range of values, and values or values of a plurality of dimensions A retrieval formula retrieval means for retrieving a registered entry group that matches the retrieval condition from the retrieval formula tree structure using a data value of given information including at least a range as a retrieval condition of the tree structure.

  In the information storage / retrieval apparatus of the present invention, data value storage means for storing data values of given information in a data value tree structure capable of simultaneously searching a range by a plurality of dimensional values is matched by a given search expression. A data value search means for searching for a data value to be received; and the apparatus accepts a registration request for the data value of the information, and uses the search expression search means for the data value. The data value storage means is used to store the data value in the data value tree structure, and the device accepts a search formula registration request and Data values may be searched from the data value tree structure using the data value search means, and the search expressions may be stored in the search expression tree structure using the search expression storage means.

  In the information storage / retrieval apparatus according to the present invention, in the determination of whether or not there is a match in the search expression search means, all dimensions included in the registered entry to be determined are searched for the search expression tree structure to be determined. All of the conditions are included, and at least the condition may be that the product set of the value range and the value range of both dimensions is not an empty set.

  In the information storage / retrieval apparatus of the present invention, the retrieval formula tree structure may be a UBI-Tree algorithm.

  The information storage search program of the present invention causes a computer to function as the search formula storage means and the search formula search means or the search formula storage means, the search formula search means, the data value storage means, and the data value search means. It is a program for.

  According to the present invention, in addition to being able to perform efficient processing by UBI-Tree for data value storage type retrieval processing as in the past, efficient processing by UBI-Tree improved for retrieval type standby retrieval processing as well. Is possible. As a result, the data value storage type search process and the search type standby type search process, which are the characteristics of uTupleSpace, can be handled without distinction, and both search processes can be efficiently realized.

The apparatus structure in this Embodiment is shown. A list of EAs held by the EA UBI-Tree (111) is shown. The structure of UBI-Tree (111) for EA is shown. The matching rule of EA and EF is shown. A list of EFs held by the EF UBI-Tree (113) is shown. The structure of UBI-Tree (113) for EF is shown. The contents of the IP address table (114) are shown. The contents of the newly registered EF are shown. The procedure of EF registration processing is shown. A list of EAs returned to the search display device (103) as a result of EF registration is shown. A list of EFs held by the EF UBI-Tree (113) after the EF registration processing is shown. The structure of UBI-Tree (113) for EF after EF registration processing is shown. The contents of the IP address table (114) after the EF registration process are shown. The content of EA newly registered is shown. The procedure of EA registration processing is shown. EA returned to the search display device (103) as a result of EA registration from the sensor device (102) is shown. A list of EAs held by the UBI-Tree (111) for EA after EA registration processing is shown. The structure of UBI-Tree (111) for EA after an EA registration process is shown. The matching rule of CA and CF is shown.

One embodiment of the present invention will be described below.
FIG. 1 shows an apparatus configuration in the present embodiment. The information storage / retrieval apparatus (101) is a server that implements uTupleSpace, and is connected to the network (104), and is an EA UBI-Tree (111) that is a data value tree structure and an EF UBI that is a search expression tree structure. -Tree (113), a request reception processing unit (115), and an IP address table (114). In addition, a sensor device (102) and a search display device (103) are connected to the network (104). The IP address of each device is 192.168.0.1 for the information storage search device (101), 192.168.0.2 for the sensor device (102), and 192.168.8.0 for the search display device (103). .3.

  The EA UBI-Tree (111) has a function of accumulating sensor values. This is a normal method of using the UBI-Tree, and is a part that could be realized by the prior art. In the following, first, the operation of this prior art will be described in detail through the EF registration process to clearly show the operation as a premise of the present invention. Next, the UBI-Tree for EF (113), which will be described in detail through the EA registration process, has a function of accumulating search expressions, which performs operations peculiar to the present invention. The differences will be clearly shown below.

  FIG. 2 shows a list of EAs held by the EA UBI-Tree (111). Here, it is assumed that, prior to the EF registration process and the EA registration process, which will be sequentially described below, five EAs (221) to (225), that is, data values of sensor information are already registered in uTupleSpace. . For example, if the meaning of these EAs is EA1 (221), “registration date and time is 11:30 on January 31, 2012, temperature is 16.2 degrees Celsius, humidity is 60%”. Indicates a measured value. For example, EA2 (222) indicates that the measured value of the sensor is “registration date and time: January 31, 2012, 13:30, temperature is 15.4 degrees Celsius”, and this data value Indicates that measurement information on humidity is not included.

  FIG. 3 shows the structure of the UBI-Tree (111) for EA. This tree structure manages the index while holding the five EAs shown in FIG. UBI-Tree has a tree structure characterized by holding multidimensional data as a data entry in a hierarchical structure. The EA UBI-Tree (111) is the same as the UBI-Tree of the prior art, and specifically performs the following operations.

  Each node of the UBI-Tree holds information of a multidimensional rectangular parallelepiped that wraps all data included in the node. For example, the node (232) holds a data entry group from EA1 to EA3. For the data entry group, the minimum value of the dimension “date and time” is “20120111130” and the maximum value is “20120111530”, the minimum value of the dimension “temperature” is “15.4”, and the maximum value is “21.8”. Yes, the minimum value of the dimension “humidity” is “35” and the maximum value is “60”. These can be thought of as representing two points in the multidimensional coordinate space that define the multidimensional cuboid, so the node (232) holds these information in tabular form. Similarly, the root node (231) holds a node (232) and a node (233), and holds information of a multidimensional rectangular parallelepiped that wraps these in a table format.

  At this time, it should be noted that even if there is a dimension that does not exist in the data entry (for example, “humidity” in EA2), it is ignored and a multidimensional rectangular parallelepiped is calculated. In addition, for the information held by the node (233), the dimension “humidity” is the symbol “φ”. This means that all data entries held by the node do not include the dimension “humidity” (the value is expressed as “φ”).

Here, an EA search process which is a search operation for the EA UBI-Tree (111), which is the prior art, is also shown. As an example, it is assumed that a search is performed with respect to this tree structure under a search condition in which “date and time” is “−∞ to + ∞” and “humidity” is “0 to 40”.
In general, a tree structure search operation is defined by recursive processing from a root node. That is, first, it is checked whether the information held by the root node (231) matches the search condition. In this case, it is checked that the product set of both value ranges is not an empty set for the dimension “date and time”, and the product set of both value ranges is not the empty set for the dimension “humidity” Since both are checked OK, it can be confirmed that this node matches the search condition.
Next, the node (232) and the node (233) are examined in the same manner recursively. In this case, since it can be confirmed that the node (232) matches the search condition, the EA1 to EA3 are sequentially examined in the same manner recursively. As a result, since it can be confirmed that only EA3 matches the search condition, it can be seen that EA3 is at least the result of the search. Note that EA2 does not include the dimension of “humidity” and therefore does not match the search.
Next, as a recursive process, the node (233) is examined. However, since this node does not include the dimension “humidity” (the value is “φ”), it does not match the search. Therefore, it is not necessary to recursively confirm EA4 to EA5 held by the node (233), and the search processing may be aborted here for this branch.
As a result of the recursive processing described above, a search result EA3 is obtained. Note that such basic processing for UBI-Tree is disclosed in detail in Patent Document 2 and Patent Document 3, including data entry insertion operations, and the implementation method is clear.

  FIG. 4 shows matching rules for EA and EF. This figure is organized in a tabular format, focusing on the handling of the case where there is no dimension, among the determination rules regarding the mismatch between the data value, that is, EA and the retrieval formula, that is, EF. . For dimensions existing in the EF (box 241), if they do not exist in the EA, they do not match (reject). If they exist, it is checked that the product set of both value ranges is not an empty set. On the other hand, dimensions that do not exist in the EF may or may not exist in the EA. Such a determination is performed on all dimensions existing in the EF, and only when all dimensions are OK, the matching between the EA and the entire EF is determined as “match”. In addition, as described above, if there is a reject during the recursive search process, it is clear that all the nodes and data entries held there are rejects. For the search operation for -Tree (111), the recursive search process may be aborted.

  FIG. 5 shows a list of EFs held by the EF UBI-Tree (113). Here, it is assumed that four EFs (251) to (254), that is, search expressions are already registered in the uTupleSpace prior to the EF registration process sequentially shown below. Note that the meaning of these EFs is, for example, EF1 (251), indicating that the registration date and time is −∞ to + ∞ and the humidity is 40% to 100%.

  FIG. 6 shows the structure of the EF UBI-Tree (113). This tree structure manages the index while holding the four EFs shown in FIG. The point here is that a value range is registered as a data entry to be registered in the UBI-Tree (unlike the UBI-Tree (111) for EA registered a value). As disclosed in Patent Document 2, UBI-Tree is an algorithm that allows not only a value but also a value range to be registered as a data entry. Basic processing of a series of tree structures for such a data entry Is clear. The insertion operation for the EF UBI-Tree (113) is the same as the operation of the conventional UBI-Tree. On the other hand, the search operation performs a characteristic operation which will be described in detail later in the present embodiment.

  FIG. 7 shows the contents of the IP address table (114). This records from which search display device each EF registered in uTupleSpace is registered. If a new EA that matches the registered EF is registered in uTupleSpace, this means that the EA is returned to the host described in this IP address table (114). Here, all the EFs are registered from the IP address “192.168.0.3” of the search display device (103).

  The information storage / retrieval apparatus (101) receives an EF registration request and an EA registration request, and performs an operation described in detail below. When there is an EF registration request, EF registration processing is performed. In the EF registration process, an EA search process (EA search operation) and an operation of inserting an EF as a search expression storage procedure are performed. When there is an EA registration request, EA registration processing is performed. In the EA registration process, an EF search process (EF search operation) is performed as a search expression search procedure, and an EA insertion operation is performed as a data value storage procedure. Any of these registration requests may be requested first, EA registration processing may be performed after EF registration processing, or EF registration processing may be performed after EA registration processing.

First, paragraphs 0033 to 0039 show a series of EF registration processes.
FIG. 8 shows the contents of the newly registered EF. Among the EFs, in particular, the value range “−∞ to + ∞” of the dimension “date and time” performs a storage type search using all EAs accumulated in the past as search targets, and stores the EFs in the future. This represents that a standby search for searching for registered EAs is performed simultaneously.

  The search display device (103) transmits the EF registration request to the information storage search device (101) via the network (104). The request acceptance processing unit (115) of the information storage / retrieval apparatus (101) accepts the registration request, and performs processing according to the following procedure.

FIG. 9 shows the procedure of the EF registration process.
First, the request reception processing unit (115) performs EA search processing as data value search means. For example, the request reception processing unit (115) performs a search operation on the EA UBI-Tree (111) using the EF as a search condition (procedure 291). Details of the search operation are as illustrated in paragraph 0027, and EA3 is obtained as a search result. Next, the search result is returned to the search display device (103) which is the request source (procedure 292).
Next, the request reception processing unit (115) compares whether or not the EF performs a standby search as a search expression storage unit and whether or not the closing price “+ ∞” of the dimension “date and time” is greater than the current time. To determine (procedure 293). Since it is true here, the request reception processing unit (115) performs an operation of inserting the EF into the EF UBI-Tree (113) as the next processing as the search expression storage unit (procedure 294). . As already shown, this insertion operation is no different from the conventional UBI-Tree.
This completes the EF registration process.

  FIG. 10 shows a list of EAs returned to the search display device (103) as a result of EF registration. Here, as already shown, the contents of EA3 are returned.

  FIG. 11 shows a list of EFs held by the EF UBI-Tree (113) after the EF registration processing. The EF registered in the EF UBI-Tree (113) in the procedure 294 is additionally held as EF5 (315) in the figure.

  FIG. 12 shows the structure of the EF UBI-Tree (113) after the EF registration process. Here, EF5 is shown added to the node (323) by the UBI-Tree algorithm.

  FIG. 13 shows the contents of the IP address table (114) after the EF registration process. This shows a state where the IP address “192.168.0.3” of the search display device (103) is added as a host corresponding to EF5. This indicates that if a new EA that matches EF5 is registered in uTupleSpace in the future, the EA will be returned to the IP address. Upon receiving the EA, the search display device (103) displays the contents on the screen.

Next, paragraphs 0040 to 0048 show a series of EA registration processes including operations extremely specific to the present invention.
FIG. 14 shows the contents of the newly registered EA. It is assumed that the sensor device (102) measures the value of “date and time” “January 31, 2012 21:30” and “temperature” “−1.5 degrees”. The sensor device (102) transmits the EA registration request to the information storage / retrieval device (101) via the network (104). The request acceptance processing unit (115) of the information storage / retrieval apparatus (101) accepts the registration request, and performs processing in the following procedure.

FIG. 15 shows the procedure of EA registration processing.
First, the request reception processing unit (115) performs EF search processing as search expression search means. For example, the request reception processing unit (115) performs a search operation on the EF UBI-Tree (113) using the EA as a search condition (procedure 351). At this time, a normal UBI-Tree registers a data value as a tree structure data entry and uses a search expression as a tree structure search condition. In this procedure, the search expression is registered as a tree structure data entry. Note that data values are used as search conditions for the tree structure. The specific procedure is as follows.

The basic search operation follows the conventional UBI-Tree. That is, first, it is checked whether the information held by the root node (321) matches the search condition. In this case, it is checked that the product set of both value ranges is not an empty set for the dimension “date and time”, and the product set of both value ranges is not the empty set for the dimension “temperature”. Since both are checked OK, it can be confirmed that this node matches the search condition.
Next, the node (322) and the node (323) are examined recursively in the same manner. In this case, the node (322) checks that the product set of the two value ranges is not an empty set with respect to the dimension of “date and time”, and is OK, but does not have the dimension of “temperature”. Is treated as a check OK. As a result, since it can be confirmed that the node (322) matches the search condition, the EF1 and EF2 are examined in order recursively. Since these are data entries, not intermediate nodes, determination is performed according to the matching rule between the original EA and EF already shown, and as a result, it can be confirmed that only EF2 matches the search condition. It turns out that it is the result of this search.
Next, as a recursive process, the node (323) is examined, and both the dimension of “date and time” and the dimension of “temperature” are checked. As a result, since it can be confirmed that the node (323) matches the search condition, EF3 to EF5 are examined in order recursively. As in the case described above, since these are data entries, determination is performed according to the original matching rule, and as a result, it can be confirmed that only EF4 matches the search condition, so EF4 is at least the result of the search. I understand that. As a result of the recursive processing described above, the request reception processing unit (115) obtains search results EF2 and EF4.

  The above rule is confirmed again by the matching rule of EA and EF shown in FIG. Since the dimension existing in the EA (box 242) is OK even if it is not included in the EF, it can be understood that the reject determination is not performed during the recursive process. Therefore, it should be noted that the recursive processing cannot be aborted in the middle, unlike the search operation for the EA UBI-Tree (111) in the prior art already shown.

  This can be easily understood by considering the example of the node (322) and EF2. Since EF2 matches all EAs that satisfy “date and time = January 31, 2012, 20: 00-22: 00”, it naturally matches the EA of FIG. 14 that is currently being registered. For this reason, in the search process, the recursive process for the lower data entry held by the node (322) cannot be omitted. Thus, based on the characteristics of the matching rule between EA and EF, the search operation of the EF UBI-Tree (113) has been modified from the normal UBI-Tree with some processing rule changes as described above. It has become.

The continuation of the procedure of the EA registration process is shown.
Next, the request reception processing unit (115) returns the search result to the corresponding search expression registration source (step 352). In this example, since EF2 and EF4 are found as search results by the procedure 351, the IP address corresponding to the EF is searched from the IP address table (114). Here, since it is understood that the IP address “192.168.0.3” is the registration source for any EF, the contents of the EA are transferred to the network (104) with respect to the IP address, that is, the search display device (103). )
Next, the request reception processing unit (115) performs an operation of inserting the EA into the EA UBI-Tree (111) as a data value storage unit (procedure 353). As already indicated, this insertion operation is a conventional UBI-Tree itself. This completes the EA registration process. In the EF registration process, the determination process (procedure 293) is necessary because the EF may or may not perform a standby search. However, in the EA registration process, all EAs are subject to accumulation. Note that no judgment processing is necessary.

  FIG. 16 shows EA returned to the search display device (103) as a result of EA registration from the sensor device (102). This is the same as the content of the EA registration request that triggered the EA registration process. Upon receiving the EA, the search display device (103) displays the contents on the screen.

  FIG. 17 shows a list of EAs held by the EA UBI-Tree (111) after the EA registration process. The registered EA is additionally held as EA6 (376) in the figure.

  FIG. 18 shows the structure of the EA UBI-Tree (111) after the EA registration process. Here, EA6 is shown added to the node (383) by the UBI-Tree algorithm. Needless to say, the EA registered in this manner is a search target in the future in another EF registration process.

  In the above, a series of processing procedures for the data value of the sensor information of the uTupleSpace, which operates correctly regardless of which of EA and EF arrives first, has been shown for EA and EF. The present invention is not limited to the processing between the EA and the EF, but can be applied as it is to the command-type communication which is another feature of the uTupleSpace, that is, the unordered matching processing between the CommandActual (CA) and the CommandForm (CF).

  FIG. 19 shows CA and CF matching rules. When this rule is compared with the matching rule of EA and EF shown in FIG. 4, it can be seen that they are exactly the same. The matching rule between CA and CF in uTupleSpace is disclosed in Patent Document 1, but the feature is that the rule of describing a single value in which and a value range in which is described is skillfully replaced. . On the other hand, there is no special point in the matching rule between CA and CF regarding the handling when the specific dimension does not exist (“φ”), which is a characteristic operation of the present invention. For this reason, the above-described series of processing procedures can be basically applied as they are by replacing “EA and EF” with “CA and CF” as they are.

  In the present invention, the UBI-Tree algorithm is used as a tree structure for indexing data entries. The scope of the present invention is not limited to this, and it is possible to search a range by a plurality of values at the same time (for example, R-Tree and its extended algorithm) and to register an entry having a value range as a data entry. (For example, it is possible to cope with a trivial extension of R-Tree) and it is possible to deal with a case where a dimension does not exist (a value of a certain dimension is “φ”) (for example, a special numerical value indicating “φ” in R-Tree). Any indexing technique that satisfies the characteristics such as registration etc. can be similarly applied. On the other hand, UBI-Tree not only satisfies the above conditions, but the data entry to be handled matches well with the requirements of the uTuple format in uTupleSpace, and the storage type search process for data entries that satisfy the above conditions is particularly efficient. Since it can be performed, the synergistic effect with the procedure exemplified in this embodiment is remarkable.

  The sensor information referred to in the present invention is not limited to the temperature and humidity information obtained by the sensor device described in the present embodiment, and can also be various information including at least other dimensions and values. Specific examples include the values measured by sensor devices, including current or voltage values, fluid flow rates, substance concentrations, brightness, noise, position, acceleration, etc. For example, it may be information acquired via the Web or the Internet. Further, in addition to these values, information including metadata indicating sensor characteristics and states, measurement date and time, and the like may be used.

  In general, when a large number of data is managed using a Tree (= tree) data structure, there is an advantage that the search speed is improved. The essential reason for improving the search speed by using the tree structure is that when inserting data entries into the tree structure, the action of “select a regular place where similar data is gathered close to each other so as to gather together” There is in doing. For example, in the binary tree algorithm that is the simplest search tree, each node has a data value and branches to two left and right descendants, and a value smaller than itself is entered on the left and a value greater than itself is entered on the right. So, determine the insertion destination. The resulting data structure is neatly arranged in ascending order when looking from the left to the right, focusing only on the X axis. In this way, since data having similar values are arranged at close positions, when searching for a target value, it can be determined easily and quickly whether the right or left branch is followed to reach the target.

  In the EA UBI-Tree and the EF UBI-Tree according to the present invention, the basic concept is the same: "data entry including similar types of items" and "data having similar values if the item types are the same. Since the insertion destination of the new data entry is determined in the direction where “entry” exists, similar values are automatically collected and managed. Therefore, in the search, candidates for which branch to follow can be narrowed down, so that a high-speed search can be performed.

  Here, the recursive processing abort operation in the EA UBI-Tree is a device for further improving the search efficiency, but the EF UBI-Tree may not be able to perform the recursive processing abort operation. The benefits of efficient search speed will not go away. In other words, the truncation (pruning) operation based on whether the range of values is out of range can be performed without any problem in any of the EA / EF UBI-Trees.

  The apparatus of the present invention can be realized by a computer and a program, and the program can be recorded on a recording medium or provided through a network.

  The present invention can be applied to the information communication industry.

101: Information storage search device 102: Sensor device 103: Search display device 104: Network 111: UBI-Tree for EA
113: UBI-Tree for EF
114: IP address table 115: request reception processing units 231, 261, 321, 381: route nodes 232, 233, 262, 263, 322, 323, 382, 383: nodes

Claims (9)

In an information accumulation and retrieval method for accumulating and retrieving information including at least a dimension and a value,
Using a search expression tree structure that can register an entry with a value range as a tree structure data entry that can be searched by a range of values of multiple dimensions at the same time. A retrieval formula accumulation procedure for accumulating the retrieved retrieval formula;
Retrieval search for searching a registered entry group that matches the search condition from the search expression tree structure using a data value of given information including at least a plurality of dimension values or value ranges as a search condition of the tree structure Procedure and
An information storage / retrieval method comprising: The information storage and retrieval method according to claim 1,
By accepting a registration request for the data value of the information and performing the search expression search procedure on the data value, a search expression can be searched from the search expression tree structure, and a range search by a multi-dimensional value can be performed simultaneously. A data value accumulation procedure for accumulating the data value in a data value tree structure before or after the retrieval formula accumulation procedure,
In the retrieval formula accumulation procedure, a data value matching the retrieval formula requested to be registered is further retrieved from the data value tree structure. The information storage and retrieval method according to claim 1 or 2,
In determining whether or not there is a match in the search formula search procedure,
All dimensions included in the registered entry of the determination target are also included in the search conditions of the search expression tree structure of the determination target, and
A method for storing and retrieving information, characterized in that at least a condition is that the product set of the value range and the value range of both dimensions is not an empty set. An information storage and retrieval method according to any one of claims 1 to 3,
The information storage retrieval method, wherein the retrieval formula tree structure is a UBI-Tree algorithm. An information storage / retrieval device for accumulating and retrieving information including at least a dimension and a value,
Using a search expression tree structure that can register an entry with a value range as a tree structure data entry that can be searched by a range of values of multiple dimensions at the same time. Search expression storage means for storing the retrieved search expressions;
Retrieval search for searching a registered entry group that matches the search condition from the search expression tree structure using a data value of given information including at least a plurality of dimension values or value ranges as a search condition of the tree structure An information storage and retrieval apparatus characterized by having means. The information storage / retrieval apparatus according to claim 5,
The device further comprises:
Data value accumulation means for accumulating data values of given information in a data value tree structure capable of range search by multiple dimensional values at the same time;
Has a data value search means to search for a matching data value by a given search expression,
further,
The device accepts a registration request for the data value of the information,
A search expression is searched from the search expression tree structure using the search expression search means for the data value, and
Using the data value storage means to store the data values in the data value tree structure;
further,
The device accepts a search type registration request,
Searching the data value from the data value tree structure using the data value search means for the search formula,
An information storage / retrieval apparatus, wherein the retrieval formula is stored in the retrieval formula tree structure using the retrieval formula storage means. The information storage / retrieval device according to claim 5 or 6,
In the determination of whether or not to match in the search expression search means,
All dimensions included in the registered entry of the determination target are also included in the search conditions of the search expression tree structure of the determination target, and
An information storage / retrieval apparatus characterized in that at least a condition that a product set of the value range and the value range of both dimensions is not an empty set is provided. An information storage / retrieval apparatus according to any one of claims 5 to 7,
The information storage and retrieval apparatus according to claim 1, wherein the retrieval formula tree structure is a UBI-Tree algorithm. A computer as the search formula storage means and the search formula search means according to any one of claims 5 to 8, or the search formula storage means, the search formula search means, the data value storage means, and the data value search means Information storage and retrieval program for functioning.

Images