Data matching device, data matching system, data matching method, and program
The data matching system addresses the issue of overlapping unique IDs by assigning order identification and using commutative hash functions, allowing secure and efficient matching of data sets with overlapping IDs, enhancing data combination security and reducing processing overhead.
Patent Information
- Application Number
- PCT/JP2023/019950
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-29
- Publication Date
- 2024-12-05
AI Technical Summary
Conventional secure matching methods fail to handle data sets with overlapping unique IDs, treating subsequent data with identical IDs as unmatched, which leads to invalid data combinations.
A data matching system that assigns order identification information to data groups with duplicate IDs, allowing for matching by generating a matching queue with unique IDs and using commutative hash functions to maintain confidentiality and enable 1:n matching.
Enables matching even when unique IDs overlap, ensuring secure data combination while maintaining confidentiality, and reducing the number of matching iterations to minimize processing time and initialization load.
Smart Images

Figure JP2023019950_05122024_PF_FP_ABST
Abstract
Description
Data matching device, data matching system, data matching method and program
[0001] The present invention relates to a data matching device, a data matching system, a data matching method, and a program.
[0002] There are various methods for combining and matching data sets between two parties using a common element (such as an ID) while maintaining confidentiality, one specific example of which is secure matching (Patent Document 1).
[0003] Secure matching has the function of combining data sets between two parties, counting the number of common elements, and outputting a summary table.
[0004] JP 2012-150276 A
[0005] In conventional secure matching, in a trial between two parties, A and B, the respective data sets must be combined 1:1 using a "unique ID" that serves as a key for data combination. Therefore, if the original data of A contains multiple (n) identical unique IDs, the second and subsequent data are deemed to have no matching destination and are treated as unmatched data (i.e., invalid data), which is a problem.
[0006] The present invention has been made in view of the above points, and has as its object to enable matching even when unique IDs overlap in one of two data sets.
[0007] In order to solve the above problem, the data matching device has an assigning unit configured to assign, for each data group in which the value of a specific element is the same among a plurality of first data sets having a first data format including the first element, identification information indicating the order in the data group to each of the first data belonging to the data group; a generating unit configured to generate subsets consisting of the first data for which there is no other first data with the same element value, the first data to which the identification information indicating the first order in each of the data groups is assigned, and subsets consisting of the first data to which the identification information indicating the second order of any of the first data from second onwards in each of the data groups is assigned; and a matching unit configured, for each of the subsets, to search for the second data to which the value of the element is the same from a plurality of second data sets having a second data format including the element, for each of the first data included in the subset.
[0008] Matching can be possible even when unique IDs overlap in one of the two data sets.
[0009] 1 is a diagram illustrating an example of a configuration of a data matching system according to an embodiment of the present invention. FIG. 1 is a diagram illustrating an example of a data set Da and a data set Db. FIG. 1 is a diagram illustrating an example of a hardware configuration of a data management device 10B according to an embodiment of the present invention. FIG. 2 is a diagram illustrating an example of a functional configuration of a data matching system according to an embodiment of the present invention. FIG. 2 is a sequence diagram illustrating an example of a processing procedure executed in the data matching system. FIG. 3 is a diagram illustrating an example of a data set Da' and a data set Db'. FIG. 4 is a flowchart illustrating an example of a processing procedure for a matching process. FIG. 4 is a diagram illustrating an example of assigning a child number to each piece of data da with duplicate "IDs" in a data set Da. FIG. 5 is a diagram illustrating an example of extraction of data da from a data set Da to a matching queue Qa. FIG. 5 is a diagram illustrating a first matching. FIG. 6 is a diagram illustrating a second matching. FIG. 7 is a diagram illustrating a third matching. FIG. 8 is a diagram illustrating a fourth matching. FIG. 9 is a diagram illustrating a process of integrating matching results.
[0010] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Fig. 1 is a diagram showing an example of the configuration of a data matching system according to an embodiment of the present invention. In Fig. 1, the data matching system includes a data management device 10A and a data management device 10B. Data management device 10A is connected to data management device 10B via a network such as the Internet.
[0011] The data management device 10A is one or more computers that store a set of multiple data items (hereinafter referred to as "data set Da") managed by organization A. The data management device 10B is one or more computers that store a set of multiple data items (hereinafter referred to as "data set Db") managed by organization B. Hereinafter, when there is no need to distinguish between the data management device 10A and the data management device 10B, they will be referred to as "data management device 10." Furthermore, when there is no need to distinguish between the data set Da and the data set Db, they will be referred to as "data set D."
[0012] Note that data set Da is an example of a plurality of data having a first data format. Data set Db is an example of a plurality of data having a second data format (different from the first data format). A data format refers to a data structure determined based on the type of elements contained in the data. An element contained in the data is the smallest unit that constitutes data, and has an element meaning (type) and an element value. In this embodiment, data set D is expressed as a data set in table format, so elements correspond to items (columns) in the table. An item name corresponds to the element meaning, and an item value corresponds to the element value. In this embodiment, an item that does not correspond to a unique ID is called an "attribute item." A unique ID is an element that is common to data set Da and data set Db, and is used to associate data when matching data set Da with data set Db.
[0013] FIG. 2 is a diagram showing examples of data sets Da and Db. In FIG. 2, data sets Da and Db are each expressed in a table format. One row in the table corresponds to one piece of data. Data set Da has an ID and a purchased item as elements (items). On the other hand, data set Db has an ID and an address as elements (items). Here, "ID" is an element common to data sets Da and Db. Therefore, in this embodiment, "ID" is a unique ID. Note that "ID" is, for example, identification information for a person. Each piece of data da in data set Da indicates one item (purchased item) purchased by a certain person. Each piece of data db in data set Db indicates the address of a certain person.
[0014] As is clear from FIG. 2 , the data set Da includes data da with duplicate IDs (the same ID). The data set Db does not have duplicate IDs (there is no data db with the same ID as another data db). In this embodiment, a data management device 10B having a data set Db with no duplicate IDs performs matching (identification) and joining between the data sets Da and Db using a common element (ID) between the data belonging to the data set Da and the data belonging to the data set Db while maintaining the confidentiality of the data sets Da and Db. Here, if the number of data da with the same ID as a certain data db is n, 1:n matching is performed for that data db.
[0015] 3 is a diagram showing an example of the hardware configuration of a data management device 10B according to an embodiment of the present invention. As shown in Fig. 3, the data management device 10B includes a drive device 100, an auxiliary storage device 102, a memory device 103, a processor 104, and an interface device 105, all of which are interconnected via a bus B.
[0016] A program that realizes processing in the data management device 10B is provided by a recording medium 101 such as a CD-ROM. When the recording medium 101 storing the program is set in the drive device 100, the program is installed from the recording medium 101 to the auxiliary storage device 102 via the drive device 100. However, the program does not necessarily have to be installed from the recording medium 101, but may be downloaded from another computer via a network. The auxiliary storage device 102 stores the installed program as well as necessary files, data, etc.
[0017] When an instruction to start a program is received, the memory device 103 reads and stores the program from the auxiliary storage device 102. The processor 104 is a CPU or a GPU (Graphics Processing Unit), or a CPU and a GPU, and executes functions related to the data management device 10B in accordance with the program stored in the memory device 103. The interface device 105 is used as an interface for connecting to a network.
[0018] The data management device 10A may also have the same hardware configuration as the data management device 10B.
[0019] FIG. 4 is a diagram showing an example of the functional configuration of a data matching system according to an embodiment of the present invention. In FIG. 4, the data management device 10A includes a preprocessing unit 11a. The preprocessing unit 11a is implemented by one or more programs installed in the data management device 10A being executed by a processor of the data management device 10A. The data management device 10A also utilizes a data storage unit 20a. The data storage unit 20a stores a data set Da. The data storage unit 20a can be implemented, for example, using an auxiliary storage device of the data management device 10A or a storage device connectable to the data management device 10A via a network.
[0020] On the other hand, the data management device 10B has a preprocessing unit 11b, a matching queue generation unit 12b, a matching unit 13b, and an integration unit 14b. Each of these units is realized by the processor 104 executing one or more programs installed in the data management device 10B. The data management device 10B also uses a data storage unit 20b. The data storage unit 20b stores a dataset Db. The data storage unit 20b can be realized, for example, using the auxiliary storage device 102 or a storage device connectable to the data management device 10B via a network.
[0021] The processing procedure executed in the data matching system will be described below. Fig. 5 is a sequence diagram for explaining an example of the processing procedure executed in the data matching system.
[0022] In step S101, the preprocessing unit 11b of the data management device 10B hashes each ID in an ID string (hereinafter referred to as "ID string b") in a dataset Db (FIG. 2) using a commutative hash function hb with an encryption key Kb. As a result, a string of hash values (hereinafter referred to as "hb (ID string b)") is generated. The ID string of the dataset Db is the string of "IDs" in the dataset Db. Next, the preprocessing unit 11b transmits hb (ID string b) to the data management device 10A (S102).
[0023] When the preprocessing unit 11a of the data management device 10A receives hb (ID string b), it hashes hb (ID string b) using a commutative hash function ha that uses the encryption key Ka (S103). As a result, a string of hash values (hereinafter referred to as "ha(hb(ID string b))") is generated.
[0024] Next, the preprocessing unit 11a hashes each ID in the ID string (hereinafter referred to as "ID string a") of the data set Da (FIG. 2) using a commutative hash function ha using the encryption key Ka (S104). As a result, a string of hash values (hereinafter referred to as "ha (ID string a)") is generated. The ID string of the data set Da is the string of "IDs" of the data set Da.
[0025] Next, the preprocessing unit 11a encrypts each attribute of the attribute string (hereinafter referred to as "attribute string a") of the data set Da (FIG. 2) using the homomorphic encryption key Kh (homomorphic encryption) to generate an attribute string a encrypted with Kh (hereinafter referred to as "Kh (attribute string a)") (S105). The attribute string of the data set Da is the string of "purchases" of the data set Da.
[0026] Next, the pre-processing unit 11a transmits to the data management device 10B (S106) an encrypted data set Da (hereinafter referred to as "data set Da'") consisting of ha (hb (ID column b)) generated in step S103, ha (ID column a) generated in step S104, and Kh (attribute column a) generated in step S105.
[0027] Upon receiving ha(hb(ID string b)) and dataset Da', the preprocessing unit 11b generates dataset Db' by replacing the ID string of dataset Db with ha(hb(ID string b)) (S107).
[0028] Next, the pre-processing unit 11b generates a string of hash values (hereinafter referred to as "hb(ha(ID string a))") by hashing ha (ID string a) of the data set Da' using a commutative hash function hb with the encryption key Kb, and replaces ha (ID string a) of the data set Da' with hb(ha(ID string a)) (S108).
[0029] 6 is a diagram showing an example of the data set Da' and the data set Db'. At the end of step S108, the data management device 10B has obtained these two data sets.
[0030] Next, the data management device 10B executes a matching process for the dataset Db' and the dataset Da' (S109).
[0031] The details of step S109 will now be described. Fig. 7 is a flowchart for explaining an example of the processing procedure of the matching process.
[0032] In step S201, for each group of data (hereinafter referred to as an "overlapping group") in which the value of the unique ID "ID" of a data set Da' overlaps (is the same as) that of other data da', the matching queue generation unit 12b assigns (unique) sequential identification information (hereinafter referred to as a "child number") within the overlapping group to each piece of data da' belonging to the overlapping group.
[0033] FIG. 8 is a diagram illustrating an example of assigning child numbers to each piece of data da' with overlapping "IDs" in a data set Da'. In the data set Da', there are multiple pieces of data da' with IDs of "005," "019," or "021." Specifically, there are two pieces of data da' with ID "005," three pieces of data da' with ID "019," and two pieces of data da' with ID "021." Therefore, the matching queue generation unit 12b assigns child numbers to each piece of data da' for each overlap group corresponding to ID "005," "019," or "021." In the example of FIG. 8, the assigned child numbers are indicated by numbers within rectangles. FIG. 8 illustrates an example in which child numbers are assigned in order of appearance from the top of the table. Note that the child numbers may be any identification information as long as the order (precedence / contiguousness) of each piece of data can be identified. However, no data within a single overlap group is assigned the same child number. That is, one piece of data is assigned to one order. Furthermore, the child numbers only need to be able to identify the order relationship within the overlapping group, and do not necessarily have to be symbols that are common across the overlapping groups. For example, the child numbers of one overlapping group may indicate the order using integers starting from 1, while the child numbers of another overlapping group may indicate the order using alphabets starting from a. Note that each ID is hashed using ha and hb, and for a given value X, hb(ha(X)) is unique.
[0034] Next, the matching queue generating unit 12b assigns 1 to a variable i (S202), which is a variable for storing the number of matchings (S204) to be described later.
[0035] Next, the matching queue generation unit 12b copies (extracts) to the matching queue Qa a subset of the data set Da' consisting of data da' with no duplicated IDs (i.e., data da' with no child number assigned) and data da' with child number i (data da' with a child number indicating that it is the i-th data). The matching queue Qa is dictionary-type (key-value type) data implemented by Map. The matching queue Qa stores the subset with IDs as keys and attributes as values. Therefore, duplicate IDs, which are keys, are not permitted.
[0036] FIG. 9 shows an example of extracting data da' from a data set Da' to a matching queue Qa. In the data set Da', data da' having unique IDs (i.e., IDs without sub-numbers) are three pieces of data da' with IDs "003," "012," and "020." There are also three pieces of data da' assigned sub-numbers indicating the first position. Therefore, in step S203, these six pieces of data da' are extracted to the matching queue Qa. The matching queue Qa generated in step S203 is reused in subsequent matching with the same size (the number of pieces of data da' that can be stored remains at six). This is to maintain security by preventing the "number of duplicate unique IDs" from being leaked (leaking the unique number of unique IDs is not considered a problem).
[0037] Next, the matching unit 13b matches the matching queue Qa with the data set Db' (FIG. 2) (S204). Specifically, for each piece of data da' extracted into the matching queue Qa, the matching unit 13b searches for data db' having the same "ID" value as the data da'. Here, due to the characteristics of a commutative hash function, for a certain value "X", hb(ha(X)) and ha(hb(X)) have the same value. Therefore, the matching unit 13b associates data da' having hb(ha(X)) as the "ID" value with data db' having ha(hb(X)) as the "ID" value.
[0038] 10 is a diagram for explaining the first matching. Fig. 10 shows a matching result Mr1 as a result of matching between the matching queue Qa and the dataset Db'. The matching result Mr1 has the same number of rows as the matching queue Qa, and has a configuration in which a column corresponding to the attribute item ("address") of the dataset Db' is added to the matching queue Qa.
[0039] The matching unit 13b determines whether or not there is data db' with the same "ID" value for each data da' in the sorting order of the matching queue Qa. For data da' for which corresponding data db' exists, the matching unit 13b copies the data da' directly to the row corresponding to the data da' in the matching result Mr1, and also copies the "address" of the data db' to the row. On the other hand, for data da' for which corresponding data db' does not exist, the matching unit 13b sets the values of all items in the row corresponding to the data da' in the matching result Mr1 to "x". Therefore, the matching result Mr1 in FIG. 10 indicates that matching was successful for data da' with IDs "ha(hb(003))," "ha(hb(005))," or "ha(hb(019))."
[0040] Next, the matching queue generation unit 12b adds 1 to i (S205). Next, the matching queue generation unit 12b extracts data da' whose child number is i into the matching queue Qa (S206). Next, the matching queue generation unit 12b extracts data da' whose child number is i from the data set Da', and copies each extracted data da' to a row in the matching queue Qa that has the same "ID" as the data da (S207). In other words, the row in the matching queue Qa is overwritten.
[0041] Next, the matching queue generation unit 12b sets a dummy value indicating that processing has been completed for the attribute item of the row corresponding to the "ID" of the data da' that is not assigned i as a child number in the data set Da' among the rows of the matching queue Qa (S207). In this embodiment, 0 is used as the dummy value. When step S207 is executed for the first time, 0 is set for the attribute item of the row corresponding to the unique "ID" (003, 012, 020).
[0042] Next, the matching queue generation unit 12b determines whether the values of the attribute items of all rows in the matching queue Qa are dummy values (S208). If there are any attribute items that are not dummy values (No in S208), step S204 and subsequent steps are repeated. If the values of the attribute items of all rows are dummy values (Yes in S208), the matching process ends.
[0043] Steps S205 and after will be described in detail. In this embodiment, when steps S206 and S207 are executed in the case where i=2, the matching queue Qa will be as shown in FIG.
[0044] FIG. 11 is a diagram illustrating the second matching. In the matching queue Qa shown in FIG. 11, the value of the attribute item corresponding to the "ID" of data da' to which i (= 2) is not assigned as a child number is set to 0. On the other hand, data da' to which i (= 2) is assigned as a child number is copied to a row in the matching queue Qa having the same "ID" as the "ID" of the data da'. In this case, since the values of the attribute items in the last three rows are not dummy values, the process returns to step S204, where matching between the matching queue Qa and the data set Db' is performed, and a matching result Mr2 is obtained. Note that the dummy values of the attribute items are also inherited in the matching result Mr2.
[0045] Next, i becomes 3 (S205). When steps S206 and S207 are executed in the case where i=3, the matching queue Qa becomes as shown in FIG.
[0046] FIG. 12 is a diagram illustrating the third matching. In the matching queue Qa shown in FIG. 12, the value of the attribute item corresponding to the "ID" of data da' to which the child number i (= 3) is not assigned is set to 0. On the other hand, data da' to which the child number i (= 3) is assigned is copied to a row in the matching queue Qa having the same "ID" as the "ID" of the data da'. In this case, since the value of the attribute item of the row with "ID" "019" is not a dummy value, the process returns to step S204, where matching between the matching queue Qa and the data set Db' is performed, and a matching result Mr3 is obtained. Note that the dummy values of the attribute items are also inherited in the matching result Mr3.
[0047] Next, i becomes 4 (S205). When steps S206 and S207 are executed in the case where i=4, the matching queue Qa becomes as shown in FIG.
[0048] 13 is a diagram for explaining the fourth matching. In the matching queue Qa shown in FIG. 12, the value of the attribute item corresponding to the "ID" of data da' to which i (=4) is not assigned as a child number is set to 0. As a result, the values of the attribute items of all data da' are set to 0. Therefore, in this case, the determination result of step S208 is Yes, and the matching process ends without performing the fourth matching. As described above, in this embodiment, three matchings are performed.
[0049] If it is revealed that the matching process is performed in order of child numbers and that matching has finished after the nth time, which is the maximum child number, the following information may be leaked: - There is at least one unique ID with a duplication count of n. - The maximum number of duplicate IDs is n.
[0050] Therefore, in use cases where confidentiality is desired, matching can be performed a predetermined number of times. For example, a threshold θ may be set in advance as an upper limit to the number of matching times, and matching may be performed the number of times equal to the threshold θ. In this case, if θ<n, matching will not be performed for data assigned child numbers θ+1 through n, but matching can be performed for other data. Furthermore, if θ>n, (θ-n) matching processes will be wasted, but the final matching result will be the same.
[0051] Returning to Fig. 5, when the matching process (S109) is completed, the integrating unit 14b executes an integration process of the matching results obtained for each matching (in the above example, the matching result Mr1 (Fig. 10), the matching result Mr2 (Fig. 11), and the matching result Mr3 (Fig. 12)) (S110).
[0052] 14 is a diagram illustrating the process of integrating matching results. As shown in FIG. 14, the process of integrating matching results generates a set of valid rows (rows that have been successfully matched) in each matching result. A valid row is a row that does not contain any "x" or dummy value (0).
[0053] Thereafter, a cross-tabulation table may be generated based on the matching results.
[0054] In the above example, the hash values of the IDs of the data sets Da and Db are the same throughout all matching (three matchings in the above example), but the hash values of the IDs of the data sets Da and Db may be changed for each matching.
[0055] In this case, if the answer is No in step S208, the process returns to step S101 in FIG. 5 instead of step S204. When step S101 and subsequent steps are executed for the second time, the encryption keys Ka and Kb used in the commutative hash function are changed. Therefore, the hash value is changed even for the same ID. As a result, duplicate IDs can be hidden, enabling more secure processing. When the processing procedure of FIG. 7 is called in step S109 for the second or subsequent time, steps S201 and S202 in FIG. 7 do not need to be executed. In step S203, the ID string in the matching queue is also replaced with a hash value based on a new encryption key. In this case, since the hash value of each ID is different from that at the time of the previous matching, it is not possible to establish a correspondence between the previous hash value and the current hash value for the same ID, and data with the same ID cannot be extracted to the same row in the matching queue Qa. Therefore, for "unique IDs (ID=003,012,020)" that only need to be processed during the first matching, the number of rows can be saved, and management can be performed so that processing can be skipped during the second and subsequent matching (i.e., by entering a dummy value in the attribute value when extracting a matching queue).Data containing duplicate IDs (such as data with child number=1 during the first matching and data with child number=2 during the second matching) can be handled during the second and subsequent matching by grouping them by ID when child numbers are assigned.
[0056] As described above, this embodiment makes it possible to match two data sets even when the unique IDs of the two data sets are the same. Therefore, of the n data sets with the same unique IDs, the second and subsequent n-1 data sets can be reflected in the counting results without being invalidated.
[0057] Furthermore, while conventional secure matching defines that "no information other than the counting result and the number of unique IDs shared by both parties (the unique number of unique IDs) will be passed on," in this embodiment, the only additional information that may be leaked is limited to the "maximum number of duplicate IDs," making it extremely secure. Moreover, depending on the settings (for example, by setting an upper limit on the number of matching attempts), this information can also be kept secret.
[0058] Furthermore, according to this embodiment, the number of matching operations performed by the matching unit 13b (the number of times step S204 in FIG. 7 is executed) can be minimized. For example, even if step S201 is not executed and data is extracted one by one from the data set Da into the matching queue Qa in step S203, no duplicate IDs will occur in the matching queue Qa, and the matching unit 13b can execute step S204. However, in this case, the number of executions of steps S203 to S207 will increase (the number of executions will be the number of data in the data set Da). As a result, the amount of initialization processing (such as terminating and resetting a launched program and copying the program to memory) required for each matching operation will increase, which is expected to result in longer processing times. However, according to this embodiment, the load of the initialization processing can be reduced.
[0059] Furthermore, according to this embodiment, by introducing the matching queue Qa and extracting a data set (subset) without overlapping IDs into the matching queue Qa, it is possible to input two data sets with no overlapping IDs to an existing matching algorithm (the matching unit 13b in this embodiment) that does not allow overlapping of unique IDs. Therefore, it is possible to reuse the existing matching algorithm and achieve 1:n matching with few development efforts.
[0060] In this embodiment, the data management device 10A is an example of a first data management device. The data management device 10B is an example of a data matching device and a second data management device. The matching queue generation unit 12b is an example of an assignment unit and a generation unit.
[0061] Although the embodiments of the present invention have been described in detail above, the present invention is not limited to such specific embodiments, and various modifications and changes are possible within the scope of the gist of the present invention described in the claims.
[0062] 10A Data management device 11a Preprocessing unit 20a Data storage unit 10B Data management device 11b Preprocessing unit 12b Matching queue generation unit 13b Matching unit 14b Integration unit 20b Data storage unit 100 Drive device 101 Recording medium 102 Auxiliary storage device 103 Memory device 104 Processor 105 Interface device B Bus
Claims
1. A data matching device comprising: an assigning unit configured to assign, for each data group among a plurality of first data sets having a first data format including a specific element, identification information of an order in the data group to each of the first data belonging to the data group in which the value of the element is the same; a generating unit configured to generate subsets consisting of the first data for which there is no other first data with the same element value and the first data to which the identification information indicating the first order in each of the data groups is assigned, and subsets consisting of the first data to which the identification information indicating the second order of any of the first data from second onwards in each of the data groups is assigned; and a matching unit configured to search for the second data having the same element value from a plurality of second data sets having a second data format including the element, for each of the first data included in the subset.
2. The data matching device according to claim 1, characterized in that the generation unit is configured to generate a subset consisting of the first data to which the identification information indicating the order has been assigned, for each order from the second order onwards in each of the data groups, until there is no more first data to which the identification information indicating the order has been assigned in any of the data groups.
3. The data matching device according to claim 1, characterized in that the matching unit is configured to generate subsets constituted by the first data to which the identification information indicating any of the same orders from the second order onwards in each of the data groups is assigned a predetermined number of times for different orders from the second order onwards.
4. A data matching system including a first data management device that stores a plurality of first data sets having a first data format including a specific element, and a second data management device that stores a plurality of second data sets having a second data format including the element, wherein the second data management device has: an assignment unit configured to assign, for each data group among the first data sets in which the element value is the same, identification information of an order in the data group to each of the first data belonging to the data group; a generation unit configured to generate subsets consisting of the first data for which there is no other first data having the same element value, and the first data to which the identification information indicating the first order in each of the data groups is assigned, and subsets consisting of the first data to which the identification information indicating the second order of any of the first data from second onwards in each of the data groups is assigned; and a matching unit configured to search for the second data having the same element value from a plurality of second data sets having a second data format including the element, for each of the first data included in the subset.
5. A data matching method characterized in that the following steps are executed by a computer: an assignment step of assigning, for each data group in which the value of a specific element is the same among a plurality of first data sets having a first data format including the specific element, identification information indicating an order in the data group to each of the first data belonging to the data group; a generation step of generating subsets constituted by the first data in which there is no other first data having the same element value, and the first data to which the identification information indicating the first order in each of the data groups is assigned, and subsets constituted by the first data to which the identification information indicating the second order of any of the first data from second onwards in each of the data groups is assigned; and a matching step of searching, for each of the first data included in the subset, for the second data having the same element value from a plurality of second data sets having a second data format including the element.
6. A program that causes a computer to execute the following steps: an assignment procedure for assigning identification information of an order in a data group to each of a plurality of first data sets having a first data format including a specific element, for each data group in which the value of the element is the same; a generation procedure for generating subsets constituted by the first data for which there is no other first data having the same element value, the first data to which the identification information indicating the first order in each of the data groups is assigned, and subsets constituted by the first data to which the identification information indicating the second order of any of the first data from second onwards in each of the data groups is assigned; and a matching procedure for searching for the second data having the same element value from a plurality of second data sets having a second data format including the element, for each of the first data included in the subset.
Citation Information
Patent Citations
Matrix / key generation device, matrix / key generation system, matrix coupling device, matrix / key generation method, and program
WO2016114309A1
Secret equijoin device, secret equijoin method, and program
WO2023276142A1
