JP6281287B2 - Information processing apparatus, search method, and program - Google Patents

Description

  The present invention relates to an information processing device, a search method, and a program.

  An example of a system that searches an attribute string using a tree structure is described in Non-Patent Document 1.

  In the multidimensional data search apparatus described in Non-Patent Document 1, multidimensional data (attribute string) that is a set of attributes is used. The search device searches a registered attribute string that matches the attribute string (search attribute string) input as a query from the registered attribute string group (registered attribute string group).

  The search device manages the registered attribute string group using a trie tree. In the trie tree, the attribute of the attribute column is associated with each hierarchy, and a node for performing matching determination of the attribute corresponding to the hierarchy is arranged. Each node has an attribute value to be compared with the search data, and a link indicating a lower layer node to be traced next when the node matches the attribute value. Furthermore, the trie tree has a leaf node that stores a registered attribute string in the lowest layer. When the search attribute string is input, the search device repeats from the root node whether or not the attribute value of each node matches the attribute value of the search attribute string until the leaf node is reached.

  Further, the search device omits (skips) the matching determination process using the attention attribute. Each time a search is performed, the search device stores a search attribute string and a node searched in a hierarchy corresponding to a predetermined attention attribute. Then, when a new search attribute string is input, the search device compares each attribute from the beginning between the past search attribute string and the new search attribute string. Here, if the attribute values up to the attribute before the attribute of interest are the same, the node for searching from the head of the new search attribute string to the attribute of interest is the same as the past search attribute string. Therefore, in this case, the search device starts searching for a new search attribute string from a node in the hierarchy corresponding to the attribute of interest.

  As described above, in the search device described in Non-Patent Document 1, the search is speeded up by skipping the match determination processing up to the attribute of interest.

Kobayashi et al., “Proposal and Evaluation of Extended Tri-Tree Method for Multi-Attribute Matching”, 2012 IEICE Communication Society Conference Proceedings 2, 2012, p. 365

  However, in the multidimensional data search device described in Non-Patent Document 1, if the attribute value before the attention attribute changes between the past search data and the new search data, the matching determination process to the attention attribute is skipped. It is not possible to obtain the effect of speeding up the search.

  An object of the present invention is to provide a search device, a search method, and a program that can solve the above-described problems and can speed up the search in searching for multidimensional data.

  An information processing apparatus according to the present invention includes an attribute column list storage unit that stores a list of attribute columns in which attribute values of a predetermined number of attributes are arranged in a predetermined attribute order, a past search attribute column, and a new search attribute column Between the position of the attribute column corresponding to the past search attribute column in the list and the position of the attribute column corresponding to the new search attribute column, calculated based on the change of the attribute value between Search means for specifying the position of the attribute string corresponding to the new search attribute string based on the distance.

  The search method of the present invention stores a list of attribute columns in which attribute values of a predetermined number of attributes are arranged in a predetermined attribute order, and sets attribute values between a past search attribute column and a new search attribute column. Based on the distance between the position of the attribute column corresponding to the past search attribute column in the list and the position of the attribute column corresponding to the new search attribute column, calculated based on the change, The position of the attribute column corresponding to the new search attribute column is specified.

  The program of the present invention stores a list of attribute columns in which attribute values of a predetermined number of attributes are arranged in a predetermined attribute order in a computer, and attributes between a past search attribute column and a new search attribute column Based on the distance between the position of the attribute column corresponding to the past search attribute column in the list and the position of the attribute column corresponding to the new search attribute column calculated based on the change in value The process of specifying the position of the attribute string corresponding to the new search attribute string is executed.

  The effect of the present invention is that the search can be speeded up in the search of multidimensional data.

It is a block diagram which shows the characteristic structure of the 1st Embodiment of this invention. It is a block diagram which shows the structure of the search system in the 1st Embodiment of this invention. It is a block diagram which shows the structure in case the search device 100 is implement | achieved by the computer in the 1st Embodiment of this invention. It is a figure which shows the example of an attribute row | line in the 1st Embodiment of this invention. It is a figure which shows the other example of an attribute row | line | column in the 1st Embodiment of this invention. It is a figure which shows the example of the attribute value information 151 in the 1st Embodiment of this invention. It is a figure which shows the example of the attribute column list 121 in the 1st Embodiment of this invention. It is a figure which shows the example of the influence value information 141 in the 1st Embodiment of this invention. It is a figure which shows the example of the search result 181 in the 1st Embodiment of this invention. It is a figure which shows the example of the difference attribute row | line | column 800 in the 1st Embodiment of this invention. It is a flowchart which shows the registration process in the 1st Embodiment of this invention. It is a flowchart which shows the list update process in the 1st Embodiment of this invention. It is a figure which shows the specific example of a registration process in the 1st Embodiment of this invention. It is a flowchart which shows the search process in the 1st Embodiment of this invention. It is a figure which shows the specific example of a search process in the 1st Embodiment of this invention. It is a figure which shows the specific example of a registration process in the 2nd Embodiment of this invention. It is a block diagram which shows the structure of the search system in the 3rd Embodiment of this invention. It is a figure which shows the example of the wild card reference information 131 in the 3rd Embodiment of this invention. It is a figure which shows the specific example of a registration process in the 3rd Embodiment of this invention. It is a figure which shows the specific example of a search process in the 3rd Embodiment of this invention.

(First embodiment)
A first embodiment of the present invention will be described.

  First, the attribute string in the first embodiment of the present invention will be described.

  In the first embodiment of the present invention, the search system selects an input attribute column (search attribute column 700 or input attribute column) from registered attribute columns (registered attribute column 600). A matching registered attribute column 600 is searched.

  4 and 5 are diagrams showing examples of attribute strings in the first embodiment of the present invention.

  In the attribute column, attribute values of a predetermined number of attributes are arranged according to a predetermined attribute order (attribute index). The attribute value may be a character or a numerical value, or may be a character, numerical value pair, set, or map. The attribute column may include information other than the attribute value for each attribute index, such as an identifier of the attribute column.

  In the attribute column of FIG. 4, the characters “a”, “a”, and “e” are associated with the attribute indexes “1”, “2”, and “3” indicating the attribute order. In the attribute column of FIG. 5, a pair of character strings {mood; good}, {motion; walking}, {position; Tokyo} is associated with the attribute indexes “1”, “2”, and “3”.

  Next, the configuration of the first exemplary embodiment of the present invention will be described.

  FIG. 2 is a block diagram showing the configuration of the search system according to the first embodiment of the present invention.

  Referring to FIG. 2, the search system according to the first exemplary embodiment of the present invention includes a search device 100, a registered attribute string input device 300, a search attribute string input device 400, and a search result output device 500. The search apparatus 100 is an embodiment of the information processing apparatus of the present invention.

  The registered attribute string input device 300 accepts the registered attribute string 600 by an input device such as a keyboard or a data file, and transmits it to the search device 100. The search attribute string input device 400 accepts the search attribute string 700 by an input device such as a keyboard or a data file, and sets a difference attribute string 800 that is a difference between the previous search attribute string 700 and the new search attribute string 700. It transmits to the search device 100. The search device 100 uses the difference attribute column 800 to search for a registered attribute column 600 that matches the new search attribute column 700. The search result output device 500 outputs (displays) the search result in the search device 100 using an output device such as a display or a data file.

  The search device 100 includes a list update unit 110, a list storage unit 120, an influence value storage unit 140, an attribute value storage unit 150, a registration unit 160, a search result storage unit 180, and a search unit 190.

  Here, the attribute value storage unit 150 stores attribute value information 151.

  The attribute value information 151 indicates a value index given to a possible value of each attribute in the attribute column.

  FIG. 6 is a diagram illustrating an example of the attribute value information 151 according to the first embodiment of this invention.

  The attribute value information 151 includes an “attribute index”, an “attribute value” indicating a possible value of the attribute of the attribute index, and a “value index” that is a value index assigned to the attribute value. In the value index, different attribute values are assigned in order from 1 to different attribute values for each attribute index.

  In the example of FIG. 6, the value indexes “0” and “1” are assigned to the attribute values “a” and “b” of the attribute index “0”, respectively. Also, value indexes “0” and “1” are assigned to the attribute values “a” and “d” of the attribute index “1”, respectively. Value indexes “0” and “1” are assigned to the attribute values “e” and “f” of the attribute index “2”, respectively.

  The list storage unit 120 stores the attribute column list 121.

  The attribute column list 121 is a list in which attribute columns related to all combinations of values that each attribute can take are arranged in ascending order of the value index.

  FIG. 7 is a diagram illustrating an example of the attribute column list 121 according to the first embodiment of this invention.

  The attribute column list 121 includes a “list index”, an “attribute column” corresponding to the list index, and “registration information”. Here, the list index indicates the position of each attribute column in the attribute column list 121. Values are assigned to the list index in order from 0 from the top attribute column of the attribute column list 121. The attribute column indicates an attribute column corresponding to the position of the list index. The registration information indicates information registered regarding the attribute column.

  In the example of FIG. 7, registration attribute columns 600 “aae” and “bdf” corresponding to the attribute columns “aae” and “bdf” are registered as registration information, respectively. In the registration information, a flag indicating whether or not the registration attribute column 600 is registered, an identifier of the registration attribute column 600, a link to other information and processing related to the registration attribute column 600 may be set. .

  The influence value storage unit 140 stores influence value information 141.

  When the attribute value in the attribute column changes between the previous and new search attribute column 700, the influence value information 141 indicates the distance between the previous and new search attribute column 700 in the attribute column list 121 (list index Indicates the magnitude of the effect on (difference).

  FIG. 8 is a diagram illustrating an example of the influence value information 141 according to the first embodiment of this invention.

  The influence value information 141 includes “attribute index” and “influence value” indicating the magnitude of the influence of the change of the attribute value of the attribute index.

  The attribute influence value Di of the attribute index i (0 ≦ i ≦ N−1, where N is the length of the attribute column) is calculated by, for example, Formula 1.

  Here, nData (j) indicates the number of values that the attribute of the attribute index j can take.

  In the example of FIG. 8, when the attribute values of the attribute indexes “0”, “1”, and “2” change between the previous and new search attribute column 700, the distance between the previous and new search attribute column 700 is as follows. , Respectively, indicating a change in multiples of “4”, “2”, and “1”.

  The list update unit 110 updates the attribute value information 151, the influence value information 141, and the attribute column list 121 when a new attribute value is included in the registered attribute column 600. Here, the list update unit 110 adds a new value index to the new attribute value in the attribute value information 151. In addition, the list update unit 110 updates the influence value information 141 by calculating the influence value using Equation 1 for each attribute. Then, the list update unit 110 inserts a new attribute column relating to all combinations of new attribute values and other attribute values into the attribute column list 121. Here, the insertion position of a new attribute column in the attribute column list 121 is calculated by the following equation (2).

  Here, V [i] indicates a value index of the attribute index i of the attribute column.

  The registration unit 160 registers the registered attribute column 600 in the attribute column list 121. Here, the registration position of the registered attribute column 600 in the attribute column list 121 is calculated by the above-described equation (2).

  The search result storage unit 180 stores the search result 181. A search result 181 indicates a search result in the previous search.

  FIG. 9 is a diagram illustrating an example of the search result 181 according to the first embodiment of this invention.

  The search result 181 includes a list index of the attribute column corresponding to the previous search attribute column 700 in the attribute column list 121. The search result 181 in FIG. 9 indicates that the list index of the attribute column corresponding to the previous search attribute column 700 is “0”.

  The search unit 190 calculates the distance between the previous search attribute column 700 and the new search attribute column 700 in the attribute column list 121 based on the difference attribute column 800 and the influence value information 141. Then, the search unit 190 acquires the list index of the attribute column corresponding to the new search attribute column 700 based on the calculated distance.

  FIG. 10 is a diagram illustrating an example of the difference attribute column 800 according to the first embodiment of this invention.

  The difference attribute column 800 indicates the previous attribute value and the new attribute value for each attribute index of the attribute that has changed between the previous search attribute column 700 and the new search attribute column 700.

  The difference attribute column 800 in FIG. 10 indicates that the value of the attribute index “0” has changed from “a” to “b”, and the attribute value of the attribute index “2” has changed from “e” to “f”.

  The search unit 190 calculates the distance between the attribute column corresponding to the previous search attribute column 700 and the attribute column corresponding to the new search attribute column 700 in the attribute column list 121 by, for example, Equation 3.

  Here, Vp [i] indicates the value index of the attribute index i of the previous search attribute column 700, and Vn [i] indicates the value index of the attribute index i of the new search attribute column 700.

  Then, the search unit 190 adds the calculated distance to the list index of the attribute column corresponding to the previous search attribute column 700 indicated by the search result 181, so that the attribute corresponding to the new search attribute column 700 is added. Get the list index of the column.

  Further, the search unit 190 acquires registration information set at the position of the list index of the attribute column corresponding to the new search attribute column 700 in the attribute column list 121, and outputs it to the search result output device 500 as a search result. To do.

  Note that the search device 100, the registered attribute string input device 300, the search attribute string input device 400, and the search result output device 500 may each be a computer that operates under control based on a program.

  FIG. 3 is a block diagram showing a configuration when the search device 100 is realized by a computer in the first embodiment of the present invention. In this case, the search device 100 includes a CPU (Central Processing Unit) 101 that executes a computer program, a storage unit (storage medium) 102 that stores data, a communication unit 103 that performs data communication with other devices, an input from a keyboard, and the like. Means 104 and output means 105 such as a display.

  The CPU 101 executes a computer program for realizing the functions of the list update unit 110, the registration unit 160, and the search unit 190. The storage unit 102 stores data of the list storage unit 120, the influence value storage unit 140, the attribute value storage unit 150, and the search result storage unit 180. The communication unit 103 receives the registration attribute string 600 and the difference attribute string 800 from the registration attribute string input device 300 and the search attribute string input device 400, respectively. In addition, the communication unit 103 transmits the search result to the search result output device 500. The input unit 104 may receive the registration attribute column 600 and the difference attribute column 800 from the user or the like, or the output unit 105 may output the search result to the user or the like.

  Next, the operation of the first exemplary embodiment of the present invention will be described.

  First, registration processing according to the first embodiment of this invention will be described.

  FIG. 11 is a flowchart showing registration processing in the first embodiment of the present invention.

  First, the search device 100 receives the registration attribute sequence 600 from the registration attribute sequence input device 300 (step S101).

  The list update unit 110 sets 0 as the target attribute index (step S102).

  The list updating unit 110 determines whether the attribute value of the target attribute index in the registered attribute column 600 is set in the attribute value information 151 (whether it is a new attribute value) (step S103).

  If it is set in the attribute value information 151 (if it is not a new attribute value) (step S103 / Y), the process proceeds to S105.

  When the attribute value information 151 is not set (in the case of a new attribute value) (step S103 / N), the list update unit 110 executes a list update process described later (step S104). As a result, the list update unit 110 registers a new attribute value in the attribute value information 151 and updates the influence value information 141. Furthermore, the list update unit 110 updates the attribute column list 121 so that the registered attribute column 600 corresponding to the attribute column including the new attribute value can be registered.

  The list update unit 110 determines whether the processing has been completed for all attribute indexes in the registered attribute column 600 (step S105).

  When processing has not been completed for all indexes (step S105 / N), the list updating unit 110 increments the target attribute index (step S106) and repeats the processing from step S103.

  When the process is completed for all indexes (step S105 / Y), the registration unit 160 calculates a list index indicating a position where the registration attribute column 600 is to be registered (step S107).

  The registration unit 160 registers the registration attribute column 600 at the calculated position of the list index (step S108).

  Next, list update processing according to the first embodiment of this invention will be described.

  FIG. 12 is a flowchart showing list update processing according to the first embodiment of this invention.

  The list update unit 110 adds the attribute value (new attribute value) of the target attribute index in the registered attribute column 600 to the attribute value information 151 (step S201). Here, the list update unit 110 adds a new value index to the new attribute value in the attribute value information 151.

  The list update unit 110 updates the influence value information 141 (step S202). Here, the list updating unit 110 calculates an influence value for each attribute using Equation (1).

  The list updating unit 110 generates a new attribute column for all combinations of the attribute value (new attribute value) of the target attribute index in the registered attribute column 600 and the attribute values that can be taken by other attribute indexes (step S203). ).

  The list update unit 110 selects one of the new attribute strings generated in step S203 (step S204).

  The list update unit 110 calculates a list index of the insertion position of the selected attribute column (step S205). Here, the list update unit 110 calculates the list index of the insertion position using Equation (2).

  The list update unit 110 inserts an attribute column at the calculated list index position in the attribute column list 121 (step S206). Here, the list update unit 110 increments the list index of each attribute column after the insertion position.

  The list update unit 110 repeats the processes of steps S204 to S206 for all the attribute strings generated in step S203 (step S207).

  Next, a specific example of the registration process according to the first embodiment of this invention will be described.

  FIG. 13 is a diagram illustrating a specific example of the registration process according to the first embodiment of this invention. Here, as in the initial state of FIG. 13, the attribute value information 151, the influence value information 141, and the attribute column list 121 are respectively converted into an attribute value storage unit 150, an influence value storage unit 140, and a list storage unit 120. It is assumed that

  Here, it is assumed that the search device 100 receives the registration attribute string 600 “bdf” as shown in FIG. In this case, the attribute value “b” of the target attribute index “0” is not set in the attribute value information 151 (it is a new attribute value). The list update unit 110 adds the attribute value “b” to the target attribute index “0” and assigns the value index “1” in the attribute value information 151 as in state 1 of FIG. The list update unit 110 calculates and sets “1”, “1”, and “1” as the influence values of the attribute indexes “0”, “1”, and “2” in the influence value information 141, respectively. The list update unit 110 generates a new attribute string “bae” including the attribute value “b” of the target attribute index “0”. The list update unit 110 calculates “1” as the list index of the attribute column “bae”. The list update unit 110 inserts the attribute string “bae” at the position of the list index “1”.

  Further, the attribute value “d” of the target attribute index “1” is not set in the attribute value information 151 (it is a new attribute value). The list update unit 110 adds the attribute value “d” to the target attribute index “1” and assigns the value index “1” in the attribute value information 151, as in state 2 in FIG. The list update unit 110 calculates and sets “2”, “1”, and “1” as the influence values of the attribute indexes “0”, “1”, and “2” in the influence value information 141, respectively. . The list update unit 110 generates new attribute strings “ade” and “bde” including the attribute value “d” of the target attribute index “1”. The list updating unit 110 calculates “1” and “3” as the list indexes of the attribute columns “ade” and “bde”, respectively. The list update unit 110 inserts attribute strings “ade” and “bde” at the positions of the list indexes “1” and “3”, respectively.

  Furthermore, the attribute value “f” of the target attribute index “2” is not set in the attribute value information 151 (it is a new attribute value). The list update unit 110 adds the attribute value “f” to the target attribute index “2” in the attribute value information 151 and assigns the value index “1” as in state 3 in FIG. The list update unit 110 calculates and sets “4”, “2”, and “1” as the influence values of the attribute indexes “0”, “1”, and “2” in the influence value information 141, respectively. The list update unit 110 generates new attribute strings “aaf”, “adf”, “baf”, “bdf” including the attribute value “f” of the target attribute index “2”. The list updating unit 110 calculates “1”, “3”, “5”, and “7” as the list indexes of the attribute columns “aaf”, “adf”, “baf”, and “bdf”, respectively. The list update unit 110 inserts attribute columns “aaf”, “adf”, “baf”, and “bdf” at the positions of the list indexes “1”, “3”, “5”, and “7”, respectively.

  Next, the registration unit 160 calculates “7” as the list index of the registration attribute column 600 “bdf”. The registration unit 160 registers the registration attribute column 600 “bdf” at the position of the list index “7” as in the state 3 in FIG.

  Next, search processing according to the first embodiment of this invention will be described.

  FIG. 14 is a flowchart showing search processing in the first embodiment of the present invention.

  First, the search device 100 receives the difference attribute sequence 800 from the search attribute sequence input device 400 (step S301).

  The search unit 190 calculates the distance between the previous search attribute column 700 and the new search attribute column 700 based on the difference attribute column 800 and the influence value information 141 (step S302). Here, the search unit 190 calculates the distance between the previous search attribute sequence 700 and the new search attribute sequence 700 using Equation (3).

  The search unit 190 calculates the list index of the attribute column corresponding to the new search attribute column 700 based on the search result 181 and the distance calculated in step S302 (step S303). Here, the search unit 190 adds the calculated distance to the list index of the attribute column corresponding to the previous search attribute column 700 indicated by the search result 181, so that the attribute corresponding to the new search attribute column 700 is added. Calculate the list index of the column.

  The search unit 190 acquires registration information registered at the position of the list index of the attribute column corresponding to the new search attribute column 700 in the attribute column list 121 (step S304). Here, for example, when the registration attribute column 600 is set as registration information, it means that the registration attribute column 600 that matches the new search attribute column 700 is registered.

  The search unit 190 outputs the acquired registration information as a search result to the search result output device 500 (step S305).

  The search unit 190 updates the search result 181 with the list index of the attribute column corresponding to the new search attribute column 700 (step S306).

  Next, a specific example of search processing according to the first embodiment of this invention will be described.

  FIG. 15 is a diagram illustrating a specific example of search processing according to the first embodiment of this invention. Here, as shown in FIG. 15, the attribute value information 151, the influence value information 141, the attribute column list 121, and the search result 181 are respectively converted into an attribute value storage unit 150, an influence value storage unit 140, a list storage unit 120, It is assumed that the search result storage unit 180 stores the search result.

  Here, it is assumed that the search apparatus 100 receives the difference attribute column 800 as shown in FIG. 15 for the previous search attribute column 700 “aae” and the new search attribute column 700 “bdf”. In this case, the search unit 190 calculates “7” as the distance between the previous search attribute column 700 “aae” and the new search attribute column 700 “bdf” as shown in FIG. “7” is calculated as the list index of the attribute column corresponding to the new search attribute column 700 “bdf”. Then, the search unit 190 acquires the registration information “bdf” registered at the position of the list index “7” and outputs it as a search result. Further, the search unit 190 updates the list index of the search result 181 with “7”.

  When the search result 181 is not stored in the search result storage unit 180, the search unit 190 compares the new search attribute column 700 with each attribute column in the attribute column list 121, for example. The list index of the attribute column corresponding to the search attribute column 700 is specified.

  Thus, the operation of the first exemplary embodiment of the present invention is completed.

  In the first embodiment of the present invention, the search device 100 receives the difference attribute sequence 800 from the search attribute sequence input device 400. However, the search device 100 may receive the new search attribute sequence 700 from the search attribute sequence input device 400 and generate the difference attribute sequence 800.

  In the first embodiment of the present invention, the search result storage unit 180 stores one search result 181. However, the search result storage unit 180 stores the search result 181 for each user identifier, and the search unit 190 uses the search result 181 corresponding to the user identifier received from the search attribute sequence input device 400 together with the difference attribute sequence 800. A search may be performed.

  In the first embodiment of the present invention, the search unit 190 uses the change in the attribute value between the previous search attribute column 700 and the new search attribute column 700 to change the previous search attribute column 700. The distance between the attribute column corresponding to と and the attribute column corresponding to the new search attribute column 700 is calculated. However, the search unit 190 may use the past search attribute sequence 700 before the previous time instead of the previous search attribute sequence 700. In this case, the search unit 190 calculates the distance between the attribute strings corresponding to the past and new search attribute strings 700 based on the change in the attribute value between the past and new search attribute strings 700. Then, the search unit 190 acquires the list index of the attribute column corresponding to the new search attribute column 700 by adding the calculated distance to the list index of the attribute column corresponding to the past search attribute column 700.

  Next, a characteristic configuration of the first exemplary embodiment of the present invention will be described. FIG. 1 is a block diagram showing a characteristic configuration of the first embodiment of the present invention.

  Referring to FIG. 1, the search device (information processing device) 100 includes a list storage unit (attribute string list storage unit) 120 and a search unit (search unit) 190. The list storage unit 120 stores an attribute column list 121 in which attribute values of a predetermined number of attributes are arranged in a predetermined attribute order. The search unit 190 is based on the distance between the attribute columns corresponding to the past and new search attribute columns in the attribute column list 121 calculated based on the change of the attribute value between the past new search attribute columns. The position of the attribute column corresponding to the new search attribute column is specified.

  According to the first embodiment of the present invention, it is possible to speed up the search in searching for multidimensional data. This is because the distance between attribute columns corresponding to the past and new search attribute columns in the attribute column list 121 calculated by the search unit 190 based on a change in attribute values between the past and new search attribute columns. This is because the position of the attribute column corresponding to the new search attribute column is specified based on the above.

  As a result, regardless of which attribute value changes between the past and the new search attribute column, the attribute column list 121 is not compared without comparing the value of each attribute between the new search attribute column and the registered attribute column. The position of the attribute column corresponding to the new search attribute column in can be specified, and the search result can be obtained.

(Second Embodiment)
Next, a second embodiment of the present invention will be described.

  In the first embodiment of the present invention, when the attribute column list 121 is updated as the registered attribute column 600 is registered, the list index of the attribute column corresponding to the previous search attribute column 700 indicated by the search result 181 Matching with the attribute column list 121 becomes impossible. Therefore, in order to obtain a correct search result, it is necessary to complete the registration of all the registered attribute columns 600 before performing a search, or to initialize the search results 181 every time the registered attribute column 600 is input. is there.

  In the second embodiment of the present invention, the search result 181 is updated along with the update of the attribute column list 121.

  The configuration of the second embodiment of the present invention is the same as that of the first embodiment (FIG. 2) of the present invention.

  The list update unit 110 updates the search result 181 when updating the attribute column list 121.

  Next, the operation of the second exemplary embodiment of the present invention will be described.

  First, a registration process according to the second embodiment of this invention will be described.

  In step S <b> 206 of the list update process (FIG. 12), the list update unit 110 performs the following process when inserting a new attribute column into the attribute column list 121. That is, the list update unit 110 increments the list index of the search result 181 when the list index of the insertion position of the new attribute column is smaller than the list index of the search result 181.

  Next, a specific example of registration processing according to the second embodiment of this invention will be described.

  FIG. 16 is a diagram showing a specific example of the registration process in the second embodiment of the present invention. Here, the attribute value information 151, the influence value information 141, and the attribute column list 121 as in the initial state of FIG. 16 are stored in the attribute value storage unit 150, the influence value storage unit 140, and the list storage unit 120, respectively. Assume that it is remembered. Further, it is assumed that the search result 181 as in the initial state of FIG. 16 is stored in the search result storage unit 180. It is assumed that

  Here, it is assumed that the search device 100 receives the registration attribute string 600 “bhf” as shown in FIG. In this case, the attribute value “h” of the target attribute index “1” is not set in the attribute value information 151 (it is a new attribute value). The list update unit 110 adds the attribute value “h” to the target attribute index “1” and assigns the value index “2” in the attribute value information 151 as in state 1 in FIG. The list updating unit 110 calculates and sets “6”, “2”, and “1” as the influence values of the attribute indexes “0”, “1”, and “2” in the influence value information 141, respectively. The list updating unit 110 generates new attribute strings “ahe”, “ahf”, “bhe”, “bhf” including the attribute value “h” of the target attribute index “1”. The list update unit 110 calculates “4” as the list index of the attribute column “ahe”. The list updating unit 110 increments the list index of the search result 181 to “8” when inserting the attribute string “ahe” at the position of the list index “4”. Furthermore, the list update unit 110 calculates “5” as the list index of the attribute column “ahf”. The list update unit 110 increments the list index of the search result 181 to “9” when the attribute string “ahf” is inserted at the position of the list index “5”.

  Note that the search processing of the second embodiment of the present invention is the same as that of the first embodiment (FIG. 14) of the present invention.

  Thus, the operation of the second exemplary embodiment of the present invention is completed.

  According to the second embodiment of the present invention, a correct search result can be obtained even if the attribute column list 121 is updated as the registered attribute column 600 is registered. The reason is that the list update unit 110 updates the search result 181 when the attribute column list 121 is updated.

(Third embodiment)
Next, a third embodiment of the present invention will be described.

  In the first embodiment of the present invention, the search cannot be performed when the attribute value of the registered attribute column 600 includes a wild card. Here, the wild card is an attribute value that matches any attribute value. In the third embodiment of the present invention, the attribute value of the wild card is indicated by “*”. For example, when the attribute value of the attribute index “1” of a certain attribute column is “*”, whatever the attribute value of the attribute index “1” of the other attribute column is, the attribute index “ It is determined that the attribute “1” matches the attribute of the attribute index “1” of the attribute column.

  In the third embodiment of the present invention, a registered attribute column 600 including a wild card attribute value is also registered in the attribute column list 121. Then, using the wild card reference information 131, a registered attribute column 600 including a wild card attribute value that matches the search attribute column 700 is searched.

  First, the configuration of the third exemplary embodiment of the present invention will be described.

  FIG. 17 is a block diagram showing a configuration of a search system according to the third embodiment of the present invention.

  Referring to FIG. 17, the search device 100 according to the third embodiment of the present invention further includes a wild card reference storage in addition to the configuration (FIG. 2) of the search device 100 according to the first embodiment of the present invention. Part 130. The wild card reference storage unit 130 stores wild card reference information 131.

  The wild card reference information 131 indicates the relationship between a certain attribute column in the attribute column list 121 and another attribute column that includes the wild card attribute value that matches the attribute column 600.

  FIG. 18 is a diagram illustrating an example of the wild card reference information 131 according to the third embodiment of this invention.

  The wild card reference information 131 includes “reference source” and “reference destination”. Here, “reference source” and “reference destination” are list indexes, and the attribute column of the list index of the reference source in the attribute column list 121 includes an attribute including the wildcard attribute value of the list index of the reference destination. Indicates that the column matches.

  In the example of FIG. 18, the reference source “0” is associated with the reference destination “4”, and the attribute column of the list index “0” includes the wild card attribute value. Indicates that it matches. That is, the attribute column registered in the list index “4” is an attribute column in which at least some of the attribute values of the attribute column of the list index “0” are replaced with wild cards.

  When a new attribute value is included in the registered attribute column 600, the list update unit 110 updates the wild card reference information 131 in addition to the attribute value information 151, the influence value information 141, and the attribute column list 121.

  Next, the operation of the third exemplary embodiment of the present invention will be described.

  First, the registration process according to the third embodiment of the present invention will be described.

  In step S206 of the above-described list update process (FIG. 12), the list update unit 110 updates the wild card reference information 131 as follows when inserting a new attribute column into the attribute column list 121. In other words, when there is another attribute column that matches the new attribute column including the wildcard attribute value in the attribute column list 121, the list update unit 110 uses the list index of the other attribute column as a reference source and creates a new attribute column. Add the list index of the attribute column to the reference destination. In addition, when there is another attribute column including a wild card attribute value that matches the new attribute column in the attribute column list 121, the list update unit 110 uses the list index of the new attribute column as the reference source, the other Add the list index of the attribute column to the reference destination.

  Here, for each attribute index of the new attribute column, the list update unit 110 creates another attribute column that matches the new attribute column including the wild card attribute value or a new attribute column by the following processing. Extract other attribute strings that contain wildcard attribute values that match.

  That is, the list update unit 110 calculates a list index obtained by subtracting the influence value for the attribute index from the list index of the insertion position of the new attribute column in the attribute column list 121. Here, when the attribute value of the attribute index of the new attribute column is a wild card, the attribute column of the calculated list index is an attribute column that matches the new attribute column. On the other hand, when the attribute value of the attribute index of the calculated list index attribute column is a wild card, the calculated attribute column of the list index is an attribute column that matches the new attribute column. The list update unit 110 repeats such subtraction of the list index and matching determination “number of attribute values for the attribute index−1” times.

  Next, a specific example of registration processing according to the third embodiment of this invention will be described.

  FIG. 19 is a diagram showing a specific example of registration processing in the third embodiment of the present invention. Here, as in the initial state of FIG. 19, the attribute value information 151, the influence value information 141, and the attribute column list 121 are respectively converted into an attribute value storage unit 150, an influence value storage unit 140, and a list storage unit 120. It is assumed that

  Here, it is assumed that the search device 100 has received a registered attribute column 600 “a * f” including a wild card attribute value as shown in FIG. In this case, the attribute value “*” of the target attribute index “1” is not set in the attribute value information 151 (it is a new attribute value). The list update unit 110 adds the attribute value “*” to the target attribute index “1” and assigns the value index “2” in the attribute value information 151 as in state 1 of FIG. The list updating unit 110 calculates and sets “6”, “2”, and “1” as the influence values of the attribute indexes “0”, “1”, and “2” in the influence value information 141, respectively. The list update unit 110 generates new attribute strings “a * e”, “a * f”, “b * e”, and “b * f” including the attribute value “*” of the target attribute index “1”. .

  The list update unit 110 calculates “4” as the list index of the attribute column “a * e”. The list update unit 110 inserts the attribute string “a * e” at the position of the list index “4”. The list update unit 110 performs a match determination on the attribute index “1” while reducing the list index in the order of “2” and “0”. Further, the list update unit 110 performs the match determination for the attribute index “2” by reducing the list index to “3”. As a result, the list update unit 110 extracts the attribute columns “ade” and “aae” of the list indexes “2” and “0” as attribute columns that match the attribute column “a * e”. Then, the list update unit 110 sets the reference destination “4” for the reference sources “0” and “2” in the wildcard reference information 131 as in the state 2 of FIG.

  Similarly, the list update unit 110 also sets the reference source, the wild card reference information 131 for the attribute columns “a * f”, “b * e”, and “b * f” as in state 2 of FIG. And the reference destination is set.

  Next, the registration unit 160 calculates “5” as the list index of the registration attribute column 600 “a * f”. The registration unit 160 registers the registration attribute column 600 “a * f” at the position of the list index “5” as in state 2 in FIG.

  Next, search processing according to the third embodiment of this invention will be described.

  In step S <b> 303 of the above-described search process (FIG. 14), the search unit 190 acquires from the wildcard reference information 131 a reference destination that uses the list index of the attribute column corresponding to the new search attribute column 700 as a reference source. Furthermore, the search unit 190 repeats acquiring a reference destination using the acquired reference destination as a reference source from the wild card reference information 131. As a result, the list index of the attribute column including the wild card that matches the new search attribute column 700 is extracted.

  In step S304, the search unit 190 acquires the list index for the new search attribute column 700 and the registration information registered at the position of the acquired reference list index. Here, for example, when the registration attribute column 600 is set as the registration information of the position of the reference list index, the registration attribute column 600 including a wild card that matches the new search attribute column 700 is registered. Means that.

  In step S305, the search unit 190 outputs the acquired registration information to the search result output device 500 as a search result.

  FIG. 20 is a diagram illustrating a specific example of search processing according to the third embodiment of the present invention. Here, as shown in FIG. 20, the attribute value information 151, the influence value information 141, the attribute column list 121, and the search result 181 are respectively converted into an attribute value storage unit 150, an influence value storage unit 140, a list storage unit 120, It is assumed that the search result storage unit 180 stores the search result. Further, it is assumed that wild card reference information 131 as illustrated in FIG. 20 is stored in the wild card reference storage unit 130.

  Here, it is assumed that the search device 100 receives the difference attribute column 800 as shown in FIG. 20 for the previous search attribute column 700 “aae” and the new search attribute column 700 “adf”. In this case, the search unit 190 calculates “3” as the distance between the previous search attribute string 700 “aae” and the new search attribute string 700 “adf” as shown in FIG. The search unit 190 calculates “3” as the list index of the attribute column corresponding to the new search attribute column 700 “adf”. Furthermore, the search unit 190 acquires a reference destination “5” for the reference source “3”. Then, the search unit 190 outputs the registration information “a * f” registered in the reference destination “5”. Further, the search unit 190 updates the list index of the search result 181 with “3”.

  Thus, the operation of the third embodiment of the present invention is completed.

  According to the third embodiment of the present invention, a search can be performed even when a wild card is included in the attribute value of the registered attribute column 600. The reason is that the search unit 190 uses the wild card reference information 131 to specify the position of the attribute column including the wild card that matches the new search attribute column 700 in the attribute column list 121.

  While the present invention has been described with reference to the embodiments, the present invention is not limited to the above embodiments. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

  The present invention can be applied to a search device or a database that searches an attribute string. Further, the present invention can be applied to a situation-dependent service such as a concierge service that requires real-time matching and real-time advertisement distribution by using user status information in an attribute column.

100 search device 101 CPU
DESCRIPTION OF SYMBOLS 102 Memory | storage means 103 Communication means 104 Input means 105 Output means 110 List update part 120 List memory | storage part 121 Attribute string list 130 Wild card reference memory | storage part 131 Wild card reference information 140 Influence value memory | storage part 141 Influence value information 150 Attribute value memory | storage part 151 Attribute value information 160 Registration unit 180 Search result storage unit 181 Search result 190 Search unit 300 Registered attribute sequence input device 400 Search attribute sequence input device 500 Search result output device 600 Registered attribute sequence 700 Search attribute sequence 800 Difference attribute sequence

Claims (10)

Attribute column list storage means for storing a list of attribute columns in which attribute values of a predetermined number of attributes are arranged in a predetermined attribute order;
The position of the attribute column corresponding to the past search attribute column in the list and the new search attribute column calculated based on the change of the attribute value between the past search attribute column and the new search attribute column Search means for specifying the position of the attribute column corresponding to the new search attribute column based on the distance between the position of the attribute column corresponding to
Equipped with a,
In each of the predetermined number of attributes, an index is assigned to each possible attribute value,
The list is a list in which attribute columns related to combinations of attribute values that can be taken by each of the predetermined number of attributes are arranged in the order of indexes.
The search means calculates the change amount of the index of the attribute value related to the attribute between the past search attribute string and the new search attribute string calculated for each of the predetermined number of attributes in the attribute string. The distance is calculated by the sum of values multiplied by the product of the number of attribute values that can be taken from the next attribute to the last attribute.
Information processing device. The search means acquires the position of the attribute column corresponding to the new search attribute column by adding the calculated distance to the position of the attribute column corresponding to the past search attribute column,
The information processing apparatus according to claim 1. Further, when a new value is added to any attribute value that can be taken by any of the predetermined number of attributes, a new index is assigned to the new value, and the attribute value that can be taken by each of the predetermined number of attributes Including a list updating means for updating the list by adding an attribute column relating to the new combination of
The information processing apparatus according to claim 1 or 2 . The list update means updates the position of the past search attribute string when the list is updated.
The information processing apparatus according to claim 3 . At least one possible attribute value of the predetermined number of attributes includes a wild card,
Furthermore, a wild card reference storage means for storing wild card reference information indicating a relationship between a position of an attribute column including a wild card in the list and a position of an attribute column matching the attribute column,
The search means specifies the position of an attribute column that includes a wildcard that matches the attribute column with respect to the position of the attribute column corresponding to the new search attribute column, based on the wild card reference information.
The information processing apparatus according to any one of claims 1 to 4. And registration means for associating information relating to the registration attribute column with a position corresponding to the registration attribute column in the list,
The search means outputs information related to the registered attribute string associated with the position of the new search attribute string;
The information processing apparatus according to any one of claims 1 to 5. An attribute column list storage means provided in the computer stores a list of attribute columns in which attribute values of a predetermined number of attributes are arranged in a predetermined attribute order;
An attribute corresponding to the past search attribute string in the list, calculated by the search means provided in the computer based on a change in attribute value between the past search attribute string and the new search attribute string A search method for identifying a position of an attribute column corresponding to the new search attribute column based on a distance between a column position and an attribute column position corresponding to the new search attribute column ,
In each of the predetermined number of attributes, an index is assigned to each possible attribute value,
The list is a list in which attribute columns related to combinations of attribute values that can be taken by each of the predetermined number of attributes are arranged in the order of indexes.
In the case where the position of the attribute column corresponding to the new search attribute column is specified, the search unit calculates the interval between the past search attribute column and the new search attribute column calculated for each of the predetermined number of attributes. The sum of the values obtained by multiplying the amount of change in the index of the attribute value related to the attribute by the product of the number of attribute values that can be taken from the attribute next to the attribute in the attribute column to the last attribute is calculated as the distance. calculate,
retrieval method. When specifying the position of the attribute string corresponding to the new search attribute string , the search means adds the calculated distance to the position of the attribute string corresponding to the past search attribute string, thereby The position of the attribute column corresponding to the search attribute column
The search method according to claim 7 . On the computer,
Storing a list of attribute columns in which attribute values of a predetermined number of attributes are arranged in a predetermined attribute order;
The position of the attribute column corresponding to the past search attribute column in the list and the new search attribute column calculated based on the change of the attribute value between the past search attribute column and the new search attribute column A program for executing a process for identifying a position of an attribute column corresponding to the new search attribute column based on a distance between the attribute column and the position of the attribute column corresponding to
In each of the predetermined number of attributes, an index is assigned to each possible attribute value,
The list is a list in which attribute columns related to combinations of attribute values that can be taken by each of the predetermined number of attributes are arranged in the order of indexes.
When specifying the position of the attribute column corresponding to the new search attribute column, it relates to the attribute between the past search attribute column and the new search attribute column calculated for each of the predetermined number of attributes. The distance is calculated by the sum of values obtained by multiplying the change amount of the attribute value index by the product of the number of attribute values that can be taken from the attribute next to the end attribute in the attribute column.
A program that executes processing.   When specifying the position of the attribute column corresponding to the new search attribute column, adding the calculated distance to the position of the attribute column corresponding to the past search attribute column, Get the position of the corresponding attribute column,
The program according to claim 9, wherein the program is executed.

Images