JP6402600B2 - Database apparatus, data management method, and program - Google Patents

Description

  The present invention relates to a column store type database device, a data management method, and a program for realizing them.

  Conventionally, as a database, a relational database that mainly processes data stored in a table in units of rows is known. The relational database is excellent in transaction processing such as update and deletion of specific data, but is not good at processing (extraction of only necessary items or aggregation processing) performed on a specific column.

  For this reason, in recent years, a database that is good at processing for each column, which is not good at relational databases, a so-called “column store database” has attracted attention (see, for example, Patent Document 1). Since the column store database holds information for each column, a specific column can be processed at high speed.

  A FAST (Filter Array STructure) structure that manages column data with a value number and a value list is known as one of the data holding structures in a column store database. FIG. 22 shows an example of a conventional FAST structure. FIG. 22A shows a table on which the FAST structure is based, and FIG. 22B shows a value number list and a value list created from the table of FIG. In the example of FIG. 22, a table having “date”, “customer ID”, and “product name” as columns is held by the FAST structure. Further, a value number list and a value list are constructed for each column.

  As shown in FIG. 22B, the value of actual data is stored in each value list in a state where duplication is eliminated. In each value number list, the number of the value list storing the value of the corresponding part is stored. In such a FAST structure, when it is desired to search for a certain value, each process is executed as follows. FIG. 23 is a diagram schematically showing processing in the FAST structure shown in FIG.

[Acquisition of list number from value list]
As shown in FIG. 23, when a value to be searched is first input, a list number of the value to be searched is acquired from the value list. Specifically, for example, when “102” is input as the customer ID, the value list number “2” is acquired in the example of FIG.

[Acquisition of number from value number list]
Next, from the value number list, a number holding the previously acquired list number (hereinafter referred to as “value number list number”) is acquired. Specifically, in the example of FIG. 23, since the acquired value list number is “2”, “2” and “5” are acquired as value number list numbers from the value number list.

  When the corresponding date and product name data are also to be searched together, the value lists stored in the value number list numbers “2” and “5” respectively from the value number lists of the date and product name respectively. Is obtained (see FIG. 22B). Next, actual data corresponding to the acquired value list number is acquired from the value list. Specifically, in this example, “2014/3/1, 102, banana” and “2014/3/3, 102, banana” are acquired.

  However, in the above-described column store database, in order to perform the above-described processing, the value list number of the actual data to be acquired and the value number list are compared by the size of the value number list. That is, if the number of data registered in the value number list is 1 billion, for example, the comparison process is repeated 1 billion times. For this reason, the column store database has a problem that the search efficiency is low and the search takes a long time. For this reason, in a column store database, an improvement in search speed is required.

  By the way, in general databases, an index is created in order to quickly find a data storage location. And in the column store database, it is considered that the search speed can be improved by creating an index. A typical example of such an index is an index adopting a B-TREE structure.

  Patent Document 2 discloses an index using a hash table. Further, Patent Document 3 discloses a system in which update processing is collectively executed by an index management device in order to reduce the time required for frequent index update processing.

JP 2013-228999 A JP 2009-140302 A JP 2011-053868 A

  However, in the case of using an index adopting the B-TREE structure, there is a problem that it is necessary to reconstruct the tree structure every time the update process is performed, and the update process takes time.

  On the other hand, when the index disclosed in Patent Document 2 is used, the problem of the index adopting the B-TREE structure can be solved. Furthermore, in this case, since the number of accesses can be reduced as compared with an index adopting the B-TREE structure, an improvement in search speed can be expected.

  However, since the index disclosed in Patent Document 2 does not store the actual data in the sorted order, if this index is applied to a column store database, the search speed may be reduced.

  In addition, even when the system disclosed in Patent Document 3 is used, the problem of the index adopting the B-TREE structure can be solved. However, in this system, it is necessary to construct an index management device, and it is necessary to construct the system. There is a problem that the cost is large.

  For this reason, in a column store database, it is required to improve the search speed without creating an index.

  An object of the present invention is to provide a database apparatus, a data management method, and a program that can improve the search speed without solving the above problems and without creating an index.

In order to achieve the above object, a database apparatus according to an aspect of the present invention provides:
A data management unit that manages the data values using a value list in which data values are registered in order of numbers, and a value number list in which the numbers of the value lists are registered in order of appearance of corresponding data values;
When a data value is specified, the number of the value list is specified for the specified data value, and further, the specified number is compared with the value number list, and the specified number in the value number list is identified. A process execution unit that extracts a registration order and executes a search process for a specified data value based on the extracted registration order;
The process execution unit
When the number of registrations in the value number list is added to each of the data values registered in the value list,
For the specified data value, the number of registrations in the value number list is specified, and when the registration rank can be extracted by the number of specified registration times, the matching between the specified number and the value number list is stopped.
It is characterized by that.

In order to achieve the above object, a data management method according to one aspect of the present invention includes:
(A) managing the data values using a value list in which data values are registered in order of numbers, and a value number list in which the numbers in the value list are registered in order of appearance of corresponding data values;
(B) When a data value is specified, the number of the value list is specified for the specified data value, and further, the specified number is compared with the value number list to determine the value number of the specified number Extracting the registration rank in the list, and executing a search process for the specified data value based on the extracted registration rank;
In the step (b),
When the number of registrations in the value number list is added to each of the data values registered in the value list,
For the specified data value, the number of registrations in the value number list is specified, and when the registration rank can be extracted by the number of specified registration times, the matching between the specified number and the value number list is stopped. It is characterized by that.

Furthermore, in order to achieve the above object, a program according to one aspect of the present invention is provided.
On the computer,
(A) managing the data values using a value list in which data values are registered in order of numbers, and a value number list in which the numbers in the value list are registered in order of appearance of corresponding data values;
(B) When a data value is specified, the number of the value list is specified for the specified data value, and further, the specified number is compared with the value number list to determine the value number of the specified number Extracting the registration rank in the list, and executing a search process for the specified data value based on the extracted registration rank,
In the step (b),
When the number of registrations in the value number list is added to each of the data values registered in the value list,
For the specified data value, the number of registrations in the value number list is specified, and when the registration rank can be extracted by the number of specified registration times, the matching between the specified number and the value number list is stopped. It is characterized by that.

  As described above, according to the present invention, it is possible to improve the search speed without creating an index.

FIG. 1 is a diagram showing an example of a data holding structure used in the present invention. FIG. 2 is an explanatory diagram for explaining search processing using the data structure shown in FIG. FIG. 3 is a block diagram showing a schematic configuration of the database apparatus according to Embodiment 1 of the present invention. FIG. 4 is a block diagram showing a specific configuration of the database apparatus according to Embodiment 1 of the present invention. FIG. 5 is a flowchart showing an operation during the list creation process of the database device according to the first embodiment of the present invention. FIG. 6 is a flowchart showing an operation during the search process of the database device in the first embodiment of the present invention. FIG. 7 is a flowchart showing an operation at the time of update processing of the database apparatus in the first embodiment of the present invention. FIG. 8 is a diagram for explaining an example of update processing according to Embodiment 1 of the present invention. FIG. 9 is a block diagram showing a specific configuration of the database apparatus according to Embodiment 2 of the present invention. FIG. 10 is a flowchart showing the operation during the search process of the database device in the second embodiment of the present invention. FIGS. 11A and 11B are diagrams for explaining search processing when there are two search target values in a conventional column store database. FIG. 12 is a diagram for explaining search processing when there are two search target values in the database apparatus according to Embodiment 2 of the present invention. FIG. 13 is a block diagram showing a specific configuration of the database apparatus according to Embodiment 2 of the present invention. FIG. 14 is a flowchart showing an operation at the time of list creation processing of the database device according to the third embodiment of the present invention. FIG. 15 is a diagram showing an example of the value number list and value list created in the third embodiment of the present invention. FIG. 16 is a flowchart showing the operation during the search process of the database device in the third embodiment of the present invention. FIG. 17 is a diagram for explaining search processing according to Embodiment 3 of the present invention. FIG. 18 is a flowchart showing the operation during the update process of the database device in the third embodiment of the present invention. FIG. 19 is a diagram for explaining the first half of the update process according to the third embodiment of the present invention. FIG. 20 is a diagram for explaining the latter half of the update processing according to Embodiment 3 of the present invention. FIG. 21 is a block diagram illustrating an example of a computer that implements the database device according to the first to third embodiments of the present invention. FIG. 22 shows an example of a conventional FAST structure. FIG. 23 is a diagram schematically showing processing in the FAST structure shown in FIG.

(Summary of Invention)
First, the outline of the present invention will be described with reference to FIGS. 1 and 2. FIG. 1 is a diagram showing an example of a data holding structure used in the present invention. FIG. 2 is an explanatory diagram for explaining search processing using the data structure shown in FIG.

  As shown in FIG. 1, the present invention also uses a column store database that employs a FAST structure as a data holding structure. In the example of FIG. 1, as in the conventional example shown in FIGS. 22A and 22B, a table having data of “date”, “customer ID”, and “product name” is a value number list and a value list. And is held by.

  Also in the example of FIG. 1, each value list has a duplicated value of actual data (hereinafter referred to as “data value”) to which a number (hereinafter referred to as “value list number”) is added. It is registered in the excluded state. In each value number list, value list numbers are registered in the order of appearance of corresponding data values.

  However, as shown in FIG. 1, in the present invention, unlike the conventional example shown in FIG. 22, an area for registering “number of registrations” in the value number list of the data value (value list number) is added to the value list. Has been. The “number of registrations” is a value obtained by counting each value list number registered in the value number list. In the example of FIG. 1, the number of items is 3, the number of rows in the value list is 3 or 4, and the number of rows in the value number list is 7. However, the present invention is not limited to this example. The present invention is applicable to any data.

  Therefore, in the present invention, at the time of searching for a value number, the search is performed until the number of target values specified by the search matches the “registration count” registered in advance. In other words, if the number of times the target value is found matches the “registration count” in the value list, it means that there are no more target values in the target value number list. The search for will end.

  For example, as illustrated in FIG. 2, when the target value is “102”, the “number of registrations” is 2, and therefore, when the value list number (2) of “102” is found twice from the value number list. The search ends. As described above, in the present invention, it is not necessary to search the value number list to the end every time, so that the search performance of the entire system can be improved.

  In the present invention, the “number of registrations” of the actual data stored in the value list is counted when the value list is created, and the initial value is set to “0”, for example. Furthermore, at the time of constructing the FAST structure, it is always necessary to check what data values exist in each item for which a value list needs to be created for all data in the original table. Therefore, since the count of the “registration count” of the data value can be executed when the value list is created, the “registration count” can be acquired without causing a heavy burden on the system.

(Embodiment 1)
Hereinafter, a database device, a data management method, and a program according to Embodiment 1 of the present invention will be described with reference to FIGS.

[Device configuration]
Initially, the structure of the database apparatus in this Embodiment 1 is demonstrated using FIG. FIG. 3 is a block diagram showing a schematic configuration of the database apparatus according to Embodiment 1 of the present invention.

  As shown in FIG. 3, the database device 100 according to the first embodiment includes a data management unit 10 and a process execution unit 30. Among these, the data management unit 10 manages data values using a value list in which data values are registered in order of numbers and a value number list in which values of the value list are registered in order of appearance of corresponding data values. (See FIG. 1).

  When the data value is designated, the process execution unit 20 first identifies the value list number for the designated data value, and further collates the identified number with the value number list to obtain the identified number. Extract registration order. Next, the process execution unit 20 executes a search process related to the specified data value based on the extracted registration order.

  In addition, when the number of registrations in the value number list for each data value registered in the value list is added to the value list, the process execution unit 20 adds the specified data value in the value number list. The number of registrations is specified (see FIG. 2). In this case, the process execution unit 20 stops collating the specified number with the value number list when the registration ranks can be extracted by the specified number of registrations (see FIG. 2).

  In this way, the database device 100 adds the number of registrations in the value number list of each data value to the value list, thereby collating the number in the value list of the specified data value with the value number list. Can be minimized. For this reason, according to the database apparatus 100, it is not necessary to always search the value number list to the end, so that the search speed can be improved without creating an index.

  Here, the configuration of the database apparatus 100 according to the first embodiment will be described more specifically with reference to FIG. FIG. 4 is a block diagram showing a specific configuration of the database apparatus according to Embodiment 1 of the present invention.

  As shown in FIG. 4, in the first embodiment, the database device 100 further includes a process reception unit 30 in addition to the data management unit 10 and the process execution unit 20. The process receiving unit 30 is connected to an external device such as a client 210 and an AP (Access Point) 220 via the network 300, and receives a process (search process, update process, etc.) requested by the external apparatus. Note that the processing requested at this time includes designation of a data value to be processed.

  In the first embodiment, the process execution unit 20 includes a process type analysis unit 21, a search processing unit 22, an update processing unit 23, and a list creation unit 24. Among these, the process type analysis unit 21 analyzes what kind of process the process received by the process reception unit 30 is, for example, a search process or an update process. Further, the processing type analysis unit 21 specifies a data value to be processed.

  In addition, the list creation unit 24 extracts data values for each column of the table from a table that registers one or more data values for each row, and creates a value list 11 and a corresponding value number list 12. create. Furthermore, the list creation unit 24 identifies the number of extractions of each data value for each extracted data value when creating the value list 11, and uses the identified number of extractions as the number of registrations of each data value. 11 is added.

  As described with reference to FIG. 1, the search processing unit 22 identifies the number (value number list number) in the value number list of the data value to be processed, and further, the value number in the identified value number list number Extract the registration order in the list. Further, the registration ranks at this time are extracted as many times as the number of registrations added to the value number list. And the search process part 22 performs a search process based on the extracted registration order. The update processing unit 23 performs the update process on the values stored in the value list 11 and the value number list 12 when the requested process is the update process.

  In the first embodiment, the data management unit 10 stores and manages data values using the FAST structure. Specifically, the data management unit 10 stores a value list 11 and a value number list 12 and manages them. In addition, in the example of FIG. 4, only one set of value list 11 and value number list 12 is illustrated, but actually, the value list 11 and value number list 12 are stored according to the number of items. Yes.

  In the value list 11, each data value, a number added to each data value (value list number), and the number of times each data value is registered are registered. In the value number list 12, a value number indicating the registration order of data values and a value list number associated with each value number are registered.

  In this embodiment, the data management unit 10 stores and manages the extraction number information 13 in addition to the value list 11 and the value number list 12. The extraction number information 13 is information that specifies the number of extractions of the registration order at the present time for the value list number to be processed. Therefore, the search processing unit 22 and the update processing unit 23 can identify the number of data values to be processed from the value number list 12 by checking the extraction number information 13.

[Device operation]
Next, the operation of the database apparatus 100 according to the first embodiment will be described with reference to FIGS. In the first embodiment, the data management method is implemented by operating the database apparatus 100. Therefore, the description of the data management method in the first embodiment is replaced with the following description of the operation of the data management apparatus 100. In the following description, FIGS. 1 to 4 are referred to as appropriate.

List creation process:
First, a process for creating a value number list and a value list will be described with reference to FIG. FIG. 5 is a flowchart showing an operation during the list creation process of the database device according to the first embodiment of the present invention. In the following, a case will be described in which the value number list and value list shown in FIG. 1 are created from the table shown in FIG.

  As shown in FIG. 5, first, the list creation unit 24 secures areas to be used for the value list 11 and the value number list 12 in the storage area of the data management unit 10, and sets 1 of items registered in the table. Is selected (step A1).

  Next, the list creation unit 24 sets the value of the parameter i indicating the number of rows in the table of the selected item to an initial value (step A2). The initial value in step A2 is “1”.

  Next, the list creation unit 24 acquires the data value of the i-th row of the item selected in step A1 from the table (see FIG. 22A) (step A3). Next, the list creation unit 24 determines whether or not the data value of the i-th line acquired in step A3 exists in the value list 11 (step A4).

  If the data value of the i-th row exists in the value list 11 as a result of the determination in step A4, the list creating unit 24 adds 1 to the registration count of this data value in the value list 11 (step A6). ).

  On the other hand, if it is determined in step A4 that the data value of the i-th row does not exist in the value list 11, the list creation unit 24 adds the data value of the i-th row to the value list 11 (step A5). Thereafter, Step A6 is executed. In step A5, the list creation unit 24 adds a value list number to the data value of the i-th row and also adds the added value list number to the value list 11.

  Next, after execution of step A6, the list creation unit 24 adds i-th as a value number in the value number list 12. Furthermore, the list creation unit 24 adds the value list number added to the data value acquired in step A3 as the value list number corresponding to the i-th (if step A5 is executed, it is added at that time) (Value list number) is added (step A7).

  Next, the list creation unit 24 adds 1 to the value of i (step A8), and then determines whether the value of i after the addition is equal to or less than the number of rows in the table (data value) (step A9). ).

  If the result of determination in step A9 is that the value of i is less than or equal to the number of rows in the table (data value), the list creation unit 24 executes step A3 again. On the other hand, if the result of determination in step A9 is that the value of i is not less than or equal to the number of rows in the table (data value), the list creation unit 24 has created the value number list 11 and the value number list 12 for all items. (Step A10).

  As a result of the determination in step A10, if the value number list 11 and the value number list 12 have not been created for all items, the list creation unit 24 executes step A1 again. On the other hand, as a result of the determination in step A10, if the value number list 11 and the value number list 12 have been created for all items, the list creation unit 24 ends the process.

  Here, the processing when “date” is selected as an item in step A1 will be specifically described (see FIG. 22A). First, if i = 1, the list creation unit 24 acquires the data value “2014/3/1” in the first row.

  Then, since “2014/3/1” does not exist in the value list, the list creation unit 24 adds the data value “2014/3/1” to the first line of the value list, and further adds this data value to this data value. A value list number “1” is added to the value. In addition, the list creation unit 24 adds 1 to the registration count of the data value “2014/3/1”.

  Furthermore, the list creation unit 24 adds “1” as the value number in the value number list 12 and further adds the value list number “1” corresponding to the value number “1”.

  Next, when i = 2, the list creation unit 24 acquires the data value “2014/3/1” in the second row. Since this data value already exists in the value list, the list creation unit 24 adds 1 to the registration count of the data value “2014/3/1” in the value list 11. In addition, the list creation unit 24 adds number 2 as the value number in the value number list 12 and further adds a value list number “1” corresponding to the value number “2”.

  Next, when i = 3, the list creation unit 24 acquires the data value “2014/3/1” in the third row. Also in this case, since the data value already exists in the value list, the list creation unit 24 adds 1 to the registration count of the data value “2014/3/1” in the value list. In addition, the list creation unit 24 adds number 3 as the value number in the value number list 12, and further adds a value list number “1” corresponding to the value number “3”.

  Next, when i = 4, the list creation unit 24 acquires the data value “2014/3/2” in the fourth row. Since the data value “2014/3/2” does not exist in the value list, the list creation unit 24 adds the data value “2014/3/2” to the second line of the value list, and further, this data value A value list number “2” is added to In addition, the list creation unit 24 adds 1 to the registration count of the data value “2014/3/2”.

  Further, the list creating unit 24 adds the number 4 as the value number in the value number list 12, and further adds the value list number “2” corresponding to the value number “4”.

  Then, the list creation unit 24 repeats the same processing until i = 7, and creates the value list and value number list shown in FIG. 1 for “date”. Further, the list creation unit 24 executes the processes from i = 1 to i = 7 for the item “customer ID” and the item “product name”, and the value list and value shown in FIG. Create a number list.

Search process:
Next, a search process using a value number list and a value list will be described with reference to FIG. FIG. 6 is a flowchart showing an operation during the search process of the database device in the first embodiment of the present invention. In the following, the case of using the value number list and value list shown in FIG. 1 will be described.

  As shown in FIG. 6, first, the search processing unit 22 acquires a value list number of a data value to be searched (hereinafter referred to as “search target value”) (step B1). Next, the search processing unit 22 acquires the number of registrations of the search target value stored in the value list (step B2).

  Next, the search processing unit 22 sets the value of the parameter j indicating the number of rows in the value number list to the initial value “1”, and further sets the value of the target data match counter k to the initial value “0”. (Step B3). The target data match counter k is a parameter indicating the number of times that the value list number of the search target value matches in the value number list. The value of the target data match counter k is stored in the data management unit 10 as the extraction number information 13.

  Next, the search processing unit 22 determines whether or not the value of k is smaller than the number of registrations of the search target value acquired in Step B2 (Step B4). As a result of the determination in step B4, when the value of k is not smaller than the number of registrations of the search target value acquired in step B2, the search processing unit 22 ends the process.

  On the other hand, if the result of determination in step B4 is that the value of k is smaller than the number of registrations of the search target value acquired in step B2, the search processing unit 22 determines that the value of j is less than the total number of values in the value number list. (Step B5).

  As a result of the determination in step B5, if the value of j is not less than the total number of rows in the value number list, the search processing unit 22 ends the process. On the other hand, if the value of j is equal to or less than the total number of rows in the value number list, the search processing unit 22 acquires a value list number associated with the value number j in the value number list (step B6).

  Next, the search processing unit 22 determines whether or not the value list number associated with the value number j acquired in step B6 matches the value list number of the search target value (step B7). .

  As a result of the determination in step B7, when the value list number associated with the value number j does not match the value list number of the search target value, the search processing unit 22 adds 1 to the value of j. (Step B8) Then, Step B6 is executed again.

  On the other hand, as a result of the determination in step B7, if the value list number associated with the value number j matches the value list number of the search target value, the search processing unit 22 returns the value number of the value number list. j is acquired as the registration order of search target values (step B9).

  Next, the search processing unit 22 adds 1 to the value of j, and further adds 1 to the value of k (step B10). Thereafter, the search processing unit 22 executes Step B4 again.

  Here, the search process when the search target value is the customer ID “102” will be specifically described (see FIG. 1). First, the search processing unit 22 acquires the value list number of “102” as a search target from the value list of the customer ID. Since the search target value “102” is stored in the second row of the value list, the search processing unit 22 acquires “2” as the value list number (see FIG. 1). Further, the search processing unit 22 acquires “2” as the registration count of the search target value “102” (see FIG. 1).

  Next, the search processing unit 22 determines whether or not the value list number associated with the value number “1” in the value number list matches the value list number of the search target value. In this case, since the former is “1” and the latter is “2”, the search processing unit 102 determines that they do not match, and adds 1 to the value of j.

  Next, the search processing unit 22 determines whether or not the value list number associated with the value number “2” in the value number list matches the value list number of the search target value. In this case, since the former is “2” and the latter is also “2”, the search processing unit 22 determines that they match. Therefore, the search processing unit 22 acquires the value number “2” in the value number list as the registration order of the search target value, and further adds 1 to each of the value of j and the value of k.

  Next, the search processing unit 22 determines whether or not the value list number associated with the value number “3” in the value number list matches the value list number of the search target value. In this case, since the former is “1” and the latter is “2”, the search processing unit 22 determines that they do not match, and adds 1 to the value of j.

  Next, the search processing unit 22 determines whether the value list number associated with the value number “4” in the value number list matches the value list number of the search target value. Also in this case, since the former is “1” and the latter is “2”, the search processing unit 102 determines that they do not match and adds 1 to the value of j.

  Next, the search processing unit 22 determines whether or not the value list number associated with the value number “5” in the value number list matches the value list number of the search target value. In this case, since the former is “2” and the latter is also “2”, the search processing unit 22 determines that they match. Therefore, the search processing unit 102 acquires the value number “5” in the value number list as the registration order of the search target value, and further adds 1 to each of the value of j and the value of k.

  As a result, the value of k becomes the same as the number of registrations of the search target value, and the determination in step B4 is No, so the search processing unit 22 ends the process.

  Further, since the value number “2” and the value number “5” are acquired as the registration order, the search processing unit 22 uses these value numbers when a search for other items is also instructed. And search the value number list and the value list of the item to be instructed. As described above, in the first embodiment, it is possible to acquire the search target value without searching the entire value number list.

Update process:
Next, the value number list and value list update processing will be described with reference to FIG. FIG. 7 is a flowchart showing an operation at the time of update processing of the database apparatus in the first embodiment of the present invention. In the following, the case of using the value number list and value list shown in FIG. 1 will be described.

  As shown in FIG. 7, first, the update processing unit 23 specifies a data value (hereinafter referred to as “pre-update data value”) that is an update target in the requested update process, and its value list. A number is acquired (step C1).

  Next, the update processing unit 23 acquires the number of registrations of the pre-update data value stored in the value list (step C2), and further subtracts 1 from the acquired number of registrations, The number of registrations stored in the value list is updated (step C3).

  Next, the update processing unit 23 specifies an updated data value (hereinafter referred to as “updated data value”) specified in the requested update process, and the specified updated data value is a value list. (Step C4).

  If the updated data value exists in the value list as a result of the determination in step C4, the update processing unit 23 acquires the number of registrations corresponding to the updated data value stored in the value list (step C5). . Next, the update processing unit 23 adds 1 to the number of registrations acquired in step C5, and updates the number of registrations stored in the value list with the obtained value (step C6).

  On the other hand, if the updated data value does not exist in the value list as a result of the determination in step C4, the update processing unit 23 stores the updated data value in the value list as a new data value (step C7). Further, the update processing unit 23 adds a new value list number to the updated data value. Next, the update processing unit 23 stores “1” as the registration count of the new data value.

  Next, after executing step C6 or C8, the update processing unit 23 stores the value list number of the updated data value in the update location of the value number list (step C9). As a result, the pre-update data value is updated.

  Next, the update processing unit 23 determines whether or not the update process has been completed for all pre-update data values (step C10). If the result of determination in step C10 is that the update processing has not been completed for all pre-update data values, the update processing unit 23 executes step C1 again. On the other hand, as a result of the determination in step C10, if the update process has been completed for all pre-update data values, the update processing unit 23 ends the process.

  Here, with reference to FIG. 8, the update process in the case where the ID of the customer who bought “lemon” in “2014/3/2” is updated to “103” will be described in detail (see FIG. 1). FIG. 8 is a diagram for explaining an example of update processing according to Embodiment 1 of the present invention.

  First, as illustrated in FIG. 8, the update processing unit 23 searches for the ID of a customer who purchased “lemon” for “2014/3/2”. In this example, in the value number list, the value number corresponding to the date “2014/3/2” is “4”.

  Therefore, since the value list number corresponding to the value number “4” in the value number list of the customer ID is “1”, the update processing unit 23 has the data value (“101” added with the value list number “1”). )) Is specified as the pre-update data value. Further, the update processing unit 23 subtracts 1 from the registration count of the pre-update data value “101” from 4 to obtain “3”.

  Next, the update processing unit 23 determines whether the updated data value “103” exists in the value list. In this example, the updated data value exists in the value list, and the value list number is “3”. Therefore, 1 is added to the corresponding registration count to obtain “2”. Further, the update processing unit 23 newly stores the value list number “3” in the update location of the value number list (value list number of the value number “4”).

  When there are a plurality of update locations, the update processing unit 23 performs the above-described processing as many times as necessary. Further, even if the number of registrations associated with the data value of the value list is changed, in the first embodiment, the existing value number list and value list can be used as they are. For this reason, in the first embodiment, unlike the case of using a conventional index, it is not necessary to recreate each list every time an update is made.

[program]
The program in the first embodiment may be a program that causes a computer to execute steps A1 to A10 shown in FIG. 5, steps B1 to B10 shown in FIG. 6, and steps C1 to C10 shown in FIG. By installing and executing this program on a computer, the database apparatus 100 and the data management method according to the first embodiment can be realized. In this case, a CPU (Central Processing Unit) of the computer functions as the process execution unit 20 and the process reception unit 30 to perform processing. The data management unit 10 can be realized by a storage device such as a hard disk provided in the computer.

[Effects of Embodiment 1]
As described above, according to the first embodiment, in the search process, it is not always necessary to search the value number list to the end, and when the search target value is found, the value number list search process ends. For this reason, the time required for the search is expected to be about half that on average compared to the conventional example, and the search speed can be improved.

  Further, in the first embodiment, when a data value is deleted, the corresponding registration count is zero. Therefore, when the deleted data value is a search target, the search process is not executed. It becomes. Therefore, an improvement in search speed is also expected from this point.

  Further, in the first embodiment, since an index is not created as in the conventional example, even if another update process is requested during the execution of the update process, the already executed update process can be continued. . Further, since no index is created, it is not necessary to regenerate the index. Therefore, since it does not take time to re-create the index, the performance of the entire system is improved.

(Embodiment 2)
Next, a database device, a data management method, and a program according to Embodiment 2 of the present invention will be described with reference to FIGS.

[Device configuration]
Initially, the structure of the database apparatus in this Embodiment 2 is demonstrated using FIG. FIG. 9 is a block diagram showing a specific configuration of the database apparatus according to Embodiment 2 of the present invention.

  As shown in FIG. 9, the database apparatus 101 in the second embodiment is similar to the database apparatus 100 shown in FIG. 2 in the first embodiment, with a data management unit 40, a process execution unit 50, and a process reception unit. 60. However, the second embodiment is different from the first embodiment in the function of the processing execution unit 50 and the information stored in the data management unit 40. Hereinafter, the difference will be specifically described.

  First, in the second embodiment, when a plurality of data values are specified, the search processing unit 52 in the processing execution unit 50 counts the number of registrations in the value number list for each specified data value (search target value). Is identified. And the search process part 52 stops collation with a value list number and a value number list | wrist at the time of having registered the registration order | rank for the number of corresponding registrations about each search object value. Further, the search processing unit 52 specifies the value list number of each search target value, creates a “test target column” that stores the specified value list number, and outputs this to the data management unit 40.

  In the second embodiment, since the search process is performed for each search target value in this way, the data management unit 40 stores and manages the extraction count information 43 by the number of search target values. The data management unit 40 also stores and manages the inspection target column information 44 that specifies the output inspection target column.

  Except for the differences described above, the database apparatus 101 in the second embodiment is configured in the same manner as the database apparatus 101 shown in FIG. 2 in the first embodiment.

[Device operation]
Next, the operation of the database apparatus 101 according to the second embodiment will be described with reference to FIG. Also in the second embodiment, the data management method is implemented by operating the database apparatus 101. Therefore, the description of the data management method in the second embodiment is also replaced with the following description of the operation of the data management apparatus 101. In the following description, FIG. 9 is referred to as appropriate.

  FIG. 10 is a flowchart showing the operation during the search process of the database device in the second embodiment of the present invention. In the following, the case of using the value number list and value list shown in FIG. 1 will be described. In the second embodiment, the list creation process and the update process are the same as those in the first embodiment, and a description thereof will be omitted.

  As shown in FIG. 10, first, the search processing unit 52 acquires a value list number for each of a plurality of search target values (step D1). In step D <b> 1, the search processing unit 52 creates a “search target column” that stores each acquired value list number, and stores this in the data management unit 40 as search target column information 44.

  Next, the search processing unit 22 acquires the number of registrations of the search target value stored in the value list for each search target value (step D2).

  Next, the search processing unit 22 sets the value of the parameter j indicating the number of rows in the value number list to the initial value “1”, and further sets the target data match counters k1, k2,... Kn (n: natural number) ) Set each value to the initial value “0” (step D3). Note that each of the target data match counters k1, k2,... Kn is a parameter indicating the number of times the value list number matches in the value number list for each search target value. Further, each of the target data matching counters k1, k2,... Kn is stored in the data management unit 10 as the extraction number information 43.

  Next, the search processing unit 52 determines whether or not all the values of k1 to kn are smaller than the number of registrations of the corresponding search target value acquired in Step D2 (Step D4). As a result of the determination in step D4, if all the values of k1 to kn are smaller than the number of registrations of the corresponding search target value, the search processing unit 52 executes step D7.

  On the other hand, as a result of the determination in step D4, when all the values of k1 to kn are not smaller than the number of registrations of the corresponding search target value (when one is more than the number of registrations), the search processing unit 52 The target data matching counter k that is equal to or greater than the registration count is specified. Then, the search processing unit 52 deletes the value list number of the search target location corresponding to the specified target data match counter k from the search target column information (step D5).

  Next, after executing step D5, the search processing unit 52 determines whether or not a value list number still exists in the search target column information (step D6). As a result of the determination in step D6, if the value list number does not yet exist in the search target column information, the search processing unit 52 ends the process. On the other hand, as a result of the determination in step D6, if the value list number still exists in the search target column information, the search processing unit 52 executes step D7.

  Next, when it is determined Yes in step D4 or when step D6 is executed, the search processing unit 52 determines whether the value of j is equal to or less than the total number of rows in the value number list ( Step D7).

  If the value of j is not less than or equal to the total number of values in the value number list as a result of the determination in step D7, the search processing unit 52 ends the process. On the other hand, if the value of j is equal to or less than the total number of rows in the value number list, the search processing unit 52 acquires the value list number associated with the value number j in the value number list (step D8).

  Next, the search processing unit 52 determines whether or not there is a value list number that matches the value list number associated with the value number j acquired in Step D8 in the search target column information (Step S8). D9).

  As a result of the determination in step D9, if there is no value list number that matches the value list number associated with the value number j, the search processing unit 52 adds 1 to the value of j (step D10). Thereafter, Step D8 is executed again.

  On the other hand, as a result of the determination in step D9, if there is a value list number that matches the value list number associated with the value number j, the search processing unit 52 searches for the value number j in the value number list. Obtained as the registration order of the target value (step D11).

  Next, the search processing unit 52 adds 1 to the value of j, further specifies the search target value having the value list number associated with the value number j, and the target data match counter of the specified search target value 1 is also added to the value of k (step D12). Thereafter, the search processing unit 52 executes Step D4 again.

  Here, a specific example in the case where two search target values are searched at once will be described with reference to FIGS. 11 and 12 (see FIG. 1). FIGS. 11A and 11B are diagrams for explaining search processing when there are two search target values in a conventional column store database. FIG. 12 is a diagram for explaining search processing when there are two search target values in the database apparatus according to Embodiment 2 of the present invention.

  As shown in FIG. 11A, it is assumed that “2014/3/1” and “2014/3/3” are input as search target values in a conventional column store database. In this case, for “2014/3/1”, first, the value list number is acquired from the value list, and the acquired value list number is compared with all value numbers in the value number list. Next, for “2014/3/3”, the acquisition of the value list number from the value list and the comparison between the acquired value list number and all the value numbers in the value number list are performed.

  In addition, as shown in FIG. 11B, in the conventional column / column store database, after obtaining the value list numbers in order for two search target values, the two value list numbers are simultaneously converted into the value number list. Can also be matched. However, even in this case, it is necessary to compare the acquired two value list numbers with all the value numbers in the value number list.

  On the other hand, as shown in FIG. 12, in the database apparatus 101 according to the second embodiment, two value list numbers of two search target values can be collated with the value number list at the same time. There is no need to always search through the list of value numbers for values. As described above, in the second embodiment, when searching for two search target values at the same time, the search range can be reduced and the search speed can be improved.

[program]
The program in the second embodiment may be a program that causes a computer to execute steps D1 to D12 shown in FIG. 10, steps A1 to A10 shown in FIG. 5, and steps C1 to C10 shown in FIG. By installing and executing this program on a computer, the database apparatus 101 and the data management method according to the second embodiment can be realized. In this case, a CPU (Central Processing Unit) of the computer functions as the process execution unit 50 and the process reception unit 60 to perform processing. The data management unit 40 can be realized by a storage device such as a hard disk provided in the computer.

(Embodiment 3)
Next, a database device, a data management method, and a program according to Embodiment 3 of the present invention will be described with reference to FIGS.

[Device configuration]
Initially, the structure of the database apparatus in this Embodiment 3 is demonstrated using FIG. FIG. 13 is a block diagram showing a specific configuration of the database apparatus according to Embodiment 2 of the present invention.

  As shown in FIG. 13, the database apparatus 102 in the third embodiment is similar to the database apparatus 100 shown in FIG. 2 in the first embodiment, with a data management unit 70, a process execution unit 80, and a process reception unit. 90. However, the third embodiment differs from the first embodiment in the function of the processing execution unit 80 and the information stored in the data management unit 70. Hereinafter, the difference will be specifically described.

  First, the list creation unit 84 in the process execution unit 80 adds the number of registrations in the value number list 71 to the value list 71 for each data value 71b registered in the value list 71 as in the first and second embodiments. . In the third embodiment, the list creation unit 84 also adds the earliest registration order (hereinafter also referred to as “start position”) in the value number list 71 of each data value (see FIG. 17 described later). .

  Further, the search processing unit 82 in the process execution unit 80 can start collating the value list number of the search target value with the value number list from the top position at the time of the search process by the start position 71d of the value list 71. it can. For this reason, according to the third embodiment, since the search start position can be narrowed, the search speed can be further improved.

  Except for the differences described above, the database apparatus 102 in the third embodiment is configured in the same manner as the database apparatus 101 shown in FIG. 2 in the first embodiment.

[Device operation]
Next, the operation of the database apparatus 102 according to the third embodiment will be described with reference to FIGS. In the first embodiment, the data management method is implemented by operating the database apparatus 100. Therefore, the description of the data management method in the first embodiment is replaced with the following description of the operation of the data management apparatus 100. In the following description, FIG. 13 is referred to as appropriate.

List creation process:
First, the creation process of the value number list and the value list will be described with reference to FIG. FIG. 14 is a flowchart showing an operation at the time of list creation processing of the database device according to the third embodiment of the present invention. In the following, a case will be described in which the value number list and value list shown in FIG. 15 are created from the table shown in FIG.

  As illustrated in FIG. 14, first, the list creation unit 84 secures areas to be used for the value list 71 and the value number list 72 in the storage area of the data management unit 70, and sets one of the items registered in the table. Is selected (step E1).

  Next, the list creation unit 84 sets the value of the parameter i indicating the number of rows in the table of the selected item to an initial value (step E2). As an initial value in step E2, “1” is given.

  Next, the list creation unit 84 acquires the data value of the i-th row of the item selected in step E1 from the table (see FIG. 22A) (step E3). Next, the list creation unit 84 determines whether or not the data value of the i-th row acquired in step E3 exists in the value list 71 (step E4).

  If the data value of the i-th row exists in the value list 71 as a result of the determination in step E4, the list creation unit 84 adds 1 to the number of registrations of this data value in the value list 71 (step E7). ).

  On the other hand, as a result of the determination in step E4, if the data value of the i-th row does not exist in the value list 71, the list creation unit 84 adds the data value of the i-th row to the value list 71 (step E5). . In step E5, the list creation unit 84 adds a value list number to the data value in the i-th row and also adds the added value list number to the value list 71.

  Subsequently, the list creation unit 84 adds i as the leading position of the data value of the i-th row in the value list 71 (step E6), and then executes step E7.

  Next, after execution of step E7, the list creation unit 84 adds the i-th number as the value number in the value number list 72. Further, the list creation unit 24 adds the value list number added to the data value acquired in step E3 as the value list number corresponding to the i-th (if step E5 is executed, the value list number is added at that time). (Value list number) is added (step E8).

  Next, the list creation unit 84 adds 1 to the value of i (step E9), and then determines whether the value of i after the addition is equal to or less than the number of rows in the table (data value) (step E10). ).

  If the value of i is equal to or less than the number of rows in the table (data value) as a result of the determination in step E10, the list creation unit 84 executes step E3 again. On the other hand, if the result of determination in step E10 is that the value of i is not less than or equal to the number of rows in the table (data value), the list creation unit 84 has created the value number list 11 and the value number list 12 for all items. Whether or not (step E11).

  As a result of the determination in step E11, if the value number list 11 and the value number list 12 have not been created for all items, the list creation unit 84 executes step E1 again. On the other hand, if it is determined in step E11 that the value number list 11 and the value number list 12 have been created for all items, the list creation unit 84 ends the process.

  Here, the processing when “date” is selected as an item in step E1 will be specifically described with reference to FIG. FIG. 15 is a diagram showing an example of the value number list and value list created in the third embodiment of the present invention. First, if i = 1, the list creation unit 24 acquires the data value “2014/3/1” in the first row.

  Then, since “2014/3/1” does not exist in the value list, the list creation unit 84 adds the data value “2014/3/1” to the first line of the value list, and further adds this data value to this data value. A value list number “1” is added to the value. The list creation unit 24 adds 1 to the number of registrations of the data value “2014/3/1” and adds “1” as the head position.

  When i = 4, the list creation unit 84 acquires the data value “2014/3/2” in the fourth row. Since the data value “2014/3/2” does not exist in the value list, the list creation unit 24 adds the data value “2014/3/2” to the second line of the value list, and further, this data value The value list number “2” and the head position “4” are added to the table. In addition, the list creation unit 24 adds 1 to the registration count of the data value “2014/3/2”.

  Similarly, when i = 5, the list creation unit 84 adds “5” as the head position of the data value “2014/3/3” in the fifth row. When i = 7, the list creation unit 84 adds “7” as the head position of the data value “2014/3/4” in the seventh row.

Search process:
Next, a search process using a value number list and a value list will be described with reference to FIG. FIG. 16 is a flowchart showing the operation during the search process of the database device in the third embodiment of the present invention. In the following, the case of using the value number list and value list shown in FIG. 15 will be described.

  As shown in FIG. 16, first, the search processing unit 82 acquires a value list number of a data value to be searched (hereinafter referred to as “search target value”) (step F1). Next, the search processing unit 82 acquires the registration count of the search target value stored in the value list (step F2). Furthermore, the search processing unit 82 also acquires the start position of the search target value stored in the value list (step F3).

  Next, the search processing unit 82 sets the value of the parameter j indicating the number of rows in the value number list to the head position acquired in step F3, and further sets the value of the target data match counter k to the initial value “0”. (Step F4). The value of the target data matching counter k is stored in the data management unit 70 as the extraction count information 73.

  Next, the search processing unit 82 determines whether or not the value of k is smaller than the number of registrations of the search target value acquired in Step F2 (Step F5). As a result of the determination in step F5, when the value of k is not smaller than the number of registrations of the search target value acquired in step F2, the search processing unit 82 ends the process.

  On the other hand, as a result of the determination in step F5, when the value of k is smaller than the number of registrations of the search target value acquired in step F2, the search processing unit 82 determines that the value of j is equal to or less than the total number of rows in the value number list. Is determined (step F6).

  As a result of the determination in step F6, when the value of j is not less than the total number of rows in the value number list, the search processing unit 82 ends the process. On the other hand, if the value of j is equal to or less than the total number of rows in the value number list, the search processing unit 82 acquires a value list number associated with the value number j in the value number list (step F7).

  Next, the search processing unit 82 determines whether the value list number associated with the value number j acquired in step F7 matches the value list number of the search target value (step F8). .

  As a result of the determination in step F8, if the value list number associated with the value number j does not match the value list number of the search target value, the search processing unit 82 adds 1 to the value of j. (Step F9) Then, Step F7 is performed again.

  On the other hand, as a result of the determination in step F8, if the value list number associated with the value number j matches the value list number of the search target value, the search processing unit 82 sets the value number in the value number list. j is acquired as the registration order of search target values (step F10).

  Next, the search processing unit 22 adds 1 to the value of j, and further adds 1 to the value of k (step F11). Thereafter, the search processing unit 22 executes Step F5 again.

  Here, the search process when the search target value is the customer ID “102” will be specifically described with reference to FIG. 17 (see FIG. 1). FIG. 17 is a diagram for explaining search processing according to Embodiment 3 of the present invention.

  As illustrated in FIG. 17, first, the search processing unit 82 acquires the value list number “102” that is the search target from the value list of the customer ID. Since the search target value “102” is stored in the second row of the value list, the search processing unit 22 acquires “2” as the value list number (see FIG. 1). Further, the search processing unit 22 acquires “2” as the registration count of the search target value “102” and acquires “2” as the head position.

  Next, since the initial value of j is set to “2”, the search processing unit 82 sets the value list number associated with the value number “2” instead of the value number “1” in the value number list. Then, it is determined whether or not the value list number of the search target value matches. In this case, since the former is “2” and the latter is also “2”, the search processing unit 22 determines that they match. Therefore, the search processing unit 22 acquires the value number “2” in the value number list as the registration order of the search target value, and further adds 1 to each of the value of j and the value of k.

  Further, even when j = 5, the search processing unit 82 matches the value number “5” in the value number list with the search target value because the value list number matches the value list number of the search target value. Obtained as the registration order of the values, and 1 is added to each of the values of j and k. As a result, the value of k becomes the same as the number of registrations of the search target value, and the determination in step E5 is No, so the search processing unit 82 ends the process.

  As described above, in the third embodiment, since the head position is used in addition to the number of registrations, the search range is further narrowed down, so that the search speed can be further improved.

Update process:
Next, the value number list and value list update processing will be described with reference to FIG. FIG. 18 is a flowchart showing the operation during the update process of the database device in the third embodiment of the present invention. In the following, the case of using the value number list and value list shown in FIG. 15 will be described.

  As shown in FIG. 18, first, the update processing unit 83 specifies the pre-update data value in the requested update process, and acquires the value list number (step G1).

  Next, the update processing unit 83 obtains the number of registrations and the start position of the pre-update data value stored in the value list (step G2), and further subtracts 1 from the obtained number of registrations. The number of registrations stored in the value list is updated with the value (step G3).

  Next, the update processing unit 83 compares the value number of the pre-update data value with the head position acquired in step G2, and determines whether or not they match (step G4). If the result of determination in step G4 is that they do not match, the update processing unit 83 executes step G7.

  On the other hand, if the result of determination in step G4 is that they match, the update processing unit 83 uses the value number list to register the same value list number as the value list number of the pre-update data value. A number is acquired (step G5). Then, the update processing unit 83 updates the leading position of the pre-update data value in the value list with the value number acquired in Step G5 (Step G6).

  Next, the update processing unit 83 specifies the updated data value specified in the requested update process, and determines whether or not the specified updated data value exists in the value list (step G7).

  If the updated data value exists in the value list as a result of the determination in step G7, the update processing unit 83 acquires the number of registrations corresponding to the updated data value stored in the value list (step G8). . Next, the update processing unit 83 adds 1 to the number of registrations acquired in step G8, and updates the number of registrations stored in the value list with the obtained value (step G9).

  Next, the update processing unit 83 determines whether or not the value number of the pre-update data value is smaller than the start position of the post-update data value in the value list in the value number list (step G10). As a result of step G10, when the value number of the updated data value is not smaller than the starting position of the updated data value in the value list, the update processing unit 83 executes step G15.

  On the other hand, as a result of step G10, when the value number of the updated data value is smaller than the starting position of the updated data value in the value list, the update processing unit 83 updates the starting position of the updated data value in the value list. Update with the value number of the subsequent data value (step G11). Thereafter, the update processing unit 83 executes Step G15.

  If the updated data value does not exist in the value list as a result of the determination in step G7, the update processing unit 83 stores the updated data value in the value list as a new data value (step G12). . Further, the update processing unit 83 adds a new value list number to the updated data value. Next, the update processing unit 83 stores “1” in the value list as the new data value registration count, and then executes Step G15.

  Next, after No determination in step G10, after execution of step G11, or after execution of step G14, the update processing unit 83 stores the value list number of the updated data value in the update location of the value number list ( Step G15). As a result, the pre-update data value is updated.

  Next, the update processing unit 83 determines whether or not the update process has been completed for all pre-update data values (step G16). If the result of determination in step G16 is that update processing has not been completed for all pre-update data values, the update processing unit 83 executes step G1 again. On the other hand, as a result of the determination in step G16, when the update process has been completed for all the pre-update data values, the update processing unit 83 ends the process.

  Here, a specific example of the update process will be described with reference to FIGS. 19 and 20 (see FIG. 1). FIG. 19 is a diagram for explaining the first half of the update process according to the third embodiment of the present invention. FIG. 20 is a diagram for explaining the latter half of the update processing according to Embodiment 3 of the present invention.

  19 and 20 show a case where the data value (customer ID) corresponding to value number 2 in the value number list is updated from “102” to “103”. As illustrated in FIG. 20, the update processing unit 83 includes the value number “2” corresponding to the pre-update data value “102” in the value number list and the pre-update data value “102” stored in the value list. Compare with the start position of.

  In this example, since both are equal, the update processing unit 83 sets the value number “5” from the value number list as the value number to which the same value list number as the value list number of the data value before update is registered next. get. Then, as illustrated in FIG. 19, the update processing unit 83 updates the leading position of the pre-update data value “102” in the value list from “2” to “5”.

  The updated data value “103” already exists in the value list, and its value list number is “3”. Therefore, as illustrated in FIG. 20, the update processing unit 83 adds 1 to the corresponding registration count to obtain “2”.

  Then, the update processing unit 83 determines whether the value number “2” of the pre-update data value is smaller than the top position of the post-update data value “103” in the value list in the value number list. In this example, the post-update data value “103” starts at “6”, and the value number of the pre-update data value is smaller. Therefore, as shown in FIG. The head position of the value “103” is updated from “6” to “2”.

  Further, as illustrated in FIG. 20, the update processing unit 23 newly stores the value list number “3” in the update location of the value number list (value list number of the value number “2”).

  Although not shown in FIGS. 19 and 20, if the updated data value does not exist in the value list, the update processing unit 83 adds the updated data value to the value list and adds a new value. Assign a value list number. In addition, the update processing unit 83 sets the number of registrations to “1” and stores the value number of the pre-update data value as the head position.

  Thus, in the third embodiment, not only when the updated data value exists in the value list but also when it does not exist, the head position is retained. Therefore, according to the third embodiment, it is not necessary to re-create an index every time it is updated as in the conventional database, so that the processing speed in the update process is also improved.

[program]
The program in the third embodiment may be any program that causes a computer to execute steps E1 to E11 shown in FIG. 14, steps F1 to F11 shown in FIG. 16, and steps G1 to G16 shown in FIG. By installing and executing this program on a computer, the database apparatus 102 and the data management method in the first embodiment can be realized. In this case, a CPU (Central Processing Unit) of the computer functions as the process execution unit 80 and the process reception unit 90 to perform processing. The data management unit 70 can be realized by a storage device such as a hard disk provided in the computer.

(Physical configuration)
Here, a computer that implements the database apparatus by executing the programs in the first to third embodiments will be described with reference to FIG. FIG. 21 is a block diagram illustrating an example of a computer that implements the database device according to the first to third embodiments of the present invention.

  As shown in FIG. 21, the computer 110 includes a CPU 111, a main memory 112, a storage device 113, an input interface 114, a display controller 115, a data reader / writer 116, and a communication interface 117. These units are connected to each other via a bus 121 so that data communication is possible.

  The CPU 111 performs various calculations by developing the program (code) in the present embodiment stored in the storage device 113 in the main memory 112 and executing them in a predetermined order. The main memory 112 is typically a volatile storage device such as a DRAM (Dynamic Random Access Memory). Further, the program in the present embodiment is provided in a state of being stored in a computer-readable recording medium 120. Note that the program in the present embodiment may be distributed on the Internet connected via the communication interface 117.

  Specific examples of the storage device 113 include a hard disk drive and a semiconductor storage device such as a flash memory. The input interface 114 mediates data transmission between the CPU 111 and an input device 118 such as a keyboard and a mouse. The display controller 115 is connected to the display device 119 and controls display on the display device 119.

  The data reader / writer 116 mediates data transmission between the CPU 111 and the recording medium 120, and reads a program from the recording medium 120 and writes a processing result in the computer 110 to the recording medium 120. The communication interface 117 mediates data transmission between the CPU 111 and another computer.

  Specific examples of the recording medium 120 include general-purpose semiconductor storage devices such as CF (Compact Flash (registered trademark)) and SD (Secure Digital), magnetic storage media such as a flexible disk, or CD- An optical storage medium such as ROM (Compact Disk Read Only Memory) can be used.

  Part or all of the above-described embodiments can be expressed by (Appendix 1) to (Appendix 12) described below, but is not limited to the following description.

(Appendix 1)
A data management unit that manages the data values using a value list in which data values are registered in order of numbers, and a value number list in which the numbers of the value lists are registered in order of appearance of corresponding data values;
When a data value is specified, the number of the value list is specified for the specified data value, and further, the specified number is compared with the value number list, and the specified number in the value number list is identified. A process execution unit that extracts a registration order and executes a search process for a specified data value based on the extracted registration order;
The process execution unit
When the number of registrations in the value number list is added to each of the data values registered in the value list,
For the specified data value, the number of registrations in the value number list is specified, and when the registration rank can be extracted by the number of specified registration times, the matching between the specified number and the value number list is stopped.
A database apparatus characterized by that.

(Appendix 2)
The process execution unit
From the table that registers one or more data values for each row, extract the data values for each column of the table, create a value list and a corresponding value number list,
Further, when creating the value list, for each of the extracted data values, the number of extractions of the data value is specified, and the extracted number is used as the registration number of the data value. Append to the list,
The database device according to attachment 1.

(Appendix 3)
When a plurality of data values are specified, the processing execution unit specifies the number of registrations in the value number list for each of the specified plurality of data values, and for each of the specified plurality of data values, When the registration rank can be extracted by the number of specified registration times, the matching between the specified number and the value number list is stopped.
The database device according to appendix 1 or 2.

(Appendix 4)
In addition to the number of registrations in the value number list for each of the data values registered in the value list, the first registration order in the value number list of the data value is added to the value list,
The database device according to any one of appendices 1 to 3, wherein the processing execution unit starts collation between the specified number and the value number list from the position of the registration order.

(Appendix 5)
(A) managing the data values using a value list in which data values are registered in order of numbers, and a value number list in which the numbers in the value list are registered in order of appearance of corresponding data values;
(B) When a data value is specified, the number of the value list is specified for the specified data value, and further, the specified number is compared with the value number list to determine the value number of the specified number Extracting the registration rank in the list, and executing a search process for the specified data value based on the extracted registration rank;
In the step (b),
When the number of registrations in the value number list is added to each of the data values registered in the value list,
For the specified data value, the number of registrations in the value number list is specified, and when the registration rank can be extracted by the number of specified registration times, the matching between the specified number and the value number list is stopped. A data management method characterized by the above.

(Appendix 6)
(C) Extracting the data value for each column of the table from a table for registering one or more data values for each row, creating a value list and a corresponding value number list,
Further, when creating the value list, for each of the extracted data values, the number of extractions of the data value is specified, and the extracted number is used as the registration number of the data value. Adding to the list, further comprising a step,
The data management method according to attachment 5.

(Appendix 7)
In the step (b), when a plurality of data values are specified, the number of registrations in the value number list is specified for each of the specified data values, and each of the specified data values is specified. When the registration rank can be extracted by the number of specified registration times, the verification of the specified number and the value number list is stopped.
The data management method according to appendix 5 or 6.

(Appendix 8)
In addition to the number of registrations in the value number list for each of the data values registered in the value list, the first registration order in the value number list of the data value is added to the value list,
The data management method according to any one of appendices 5 to 7, wherein, in the step (b), collation between the specified number and the value number list is started from the position of the registration order.

(Appendix 9)
On the computer,
(A) managing the data values using a value list in which data values are registered in order of numbers, and a value number list in which the numbers in the value list are registered in order of appearance of corresponding data values;
(B) When a data value is specified, the number of the value list is specified for the specified data value, and further, the specified number is compared with the value number list to determine the value number of the specified number Extracting the registration rank in the list, and executing a search process for the specified data value based on the extracted registration rank,
In the step (b),
When the number of registrations in the value number list is added to each of the data values registered in the value list,
For the specified data value, the number of registrations in the value number list is specified, and when the registration rank can be extracted by the number of specified registration times, the matching between the specified number and the value number list is stopped. program.

(Appendix 10)
In the computer,
(C) Extracting the data value for each column of the table from a table for registering one or more data values for each row, creating a value list and a corresponding value number list,
Further, when creating the value list, for each of the extracted data values, the number of extractions of the data value is specified, and the extracted number is used as the registration number of the data value. Add to the list, further execute steps,
The program according to appendix 9.

(Appendix 11)
In the step (b), when a plurality of data values are specified, the number of registrations in the value number list is specified for each of the specified data values, and each of the specified data values is specified. When the registration rank can be extracted by the number of specified registration times, the verification of the specified number and the value number list is stopped.
The program according to appendix 9 or 10.

(Appendix 12)
In addition to the number of registrations in the value number list for each of the data values registered in the value list, the first registration order in the value number list of the data value is added to the value list,
The program according to any one of appendices 9 to 11, wherein in the step (b), collation between the specified number and the value number list is started from the position of the registration order.

    As described above, according to the present invention, it is possible to improve the search speed without creating an index. The present invention is useful for databases, particularly column store type databases.

10, 40, 70 Data management unit 11, 41, 71 Value list 12, 42, 72 Value number list 13, 43, 73 Extraction count information 20, 50, 80 Processing execution unit 21, 51, 81 Processing type analysis unit 22, 52, 82 Search processing unit 23, 53, 83 Update processing unit 24, 54, 84 List creation unit 30, 60, 90 Processing reception unit 44 Search target column information 100 Database device (Embodiment 1)
101 Database device (Embodiment 2)
103 Database device (Embodiment 3)
110 Computer 111 CPU
112 Main Memory 113 Storage Device 114 Input Interface 115 Display Controller 116 Data Reader / Writer 117 Communication Interface 118 Input Device 119 Display Device 120 Recording Medium 121 Bus

Claims (6)

A data management unit that manages the data values using a value list in which data values are registered in order of numbers, and a value number list in which the numbers of the value lists are registered in order of appearance of corresponding data values;
When a data value is specified, the number of the value list is specified for the specified data value, and further, the specified number is compared with the value number list, and the specified number in the value number list is identified. A process execution unit that extracts a registration order and executes a search process for a specified data value based on the extracted registration order;
The process execution unit
When the number of registrations in the value number list is added to each of the data values registered in the value list,
For the specified data value, the number of registrations in the value number list is specified, and when the registration rank can be extracted by the number of specified registration times, the matching between the specified number and the value number list is stopped.
A database apparatus characterized by that. The process execution unit
From the table that registers one or more data values for each row, extract the data values for each column of the table, create a value list and a corresponding value number list,
Further, when creating the value list, for each of the extracted data values, the number of extractions of the data value is specified, and the extracted number is used as the registration number of the data value. Append to the list,
The database device according to claim 1. When a plurality of data values are specified, the processing execution unit specifies the number of registrations in the value number list for each of the specified plurality of data values, and for each of the specified plurality of data values, When the registration rank can be extracted by the number of specified registration times, the matching between the specified number and the value number list is stopped.
The database apparatus according to claim 1 or 2. In addition to the number of registrations in the value number list for each of the data values registered in the value list, the first registration order in the value number list of the data value is added to the value list,
The database apparatus according to any one of claims 1 to 3, wherein the processing execution unit starts collation between the identified number and the value number list from the position of the registration order. (A) managing the data values using a value list in which data values are registered in order of numbers, and a value number list in which the numbers in the value list are registered in order of appearance of corresponding data values;
(B) When a data value is specified, the number of the value list is specified for the specified data value, and further, the specified number is compared with the value number list to determine the value number of the specified number Extracting the registration rank in the list, and executing a search process for the specified data value based on the extracted registration rank;
In the step (b),
When the number of registrations in the value number list is added to each of the data values registered in the value list,
For the specified data value, the number of registrations in the value number list is specified, and when the registration rank can be extracted by the number of specified registration times, the matching between the specified number and the value number list is stopped. A data management method characterized by the above. On the computer,
(A) managing the data values using a value list in which data values are registered in order of numbers, and a value number list in which the numbers in the value list are registered in order of appearance of corresponding data values;
(B) When a data value is specified, the number of the value list is specified for the specified data value, and further, the specified number is compared with the value number list to determine the value number of the specified number Extracting the registration rank in the list, and executing a search process for the specified data value based on the extracted registration rank,
In the step (b),
When the number of registrations in the value number list is added to each of the data values registered in the value list,
For the specified data value, the number of registrations in the value number list is specified, and when the registration rank can be extracted by the number of specified registration times, the matching between the specified number and the value number list is stopped. program.

Images