JP2000066933A - Time series data management system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To allow the latest data and time series data of plural past points to coexist in the same table. SOLUTION: A time series table definition means 1 generates a validity term index 6 defining end time and start time and a unique index 5 including the definition of the end time. A time series table updating means 8 controls the preparation of time series data 4 based on a value at stored time and the insertion of the data 4 into a time series table 2 at the time of receiving an insert/update/delete instruction for latest data, executes the insert/update/delete of the latest data 3 and updates both the indexes 5, 6. A time series table retrieving means 12 considers retrieving time for a retrieving instruction adds 'a condition for selecting only intra-table data at specifiic time' to a retrieving condition and retrieves the time series table 2 under the retrieving condition by utilizing both the indexes 5, 6.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION The present invention relates to time-series data (arbitrary plural time points) that is effective for statistical work and the like performed based on the contents at any point in a relational database in an online information processing system having a relational database as a core. Of the relational database) and a time-series data management method having a search function.

[0002]

2. Description of the Related Art FIG. 14 is a block diagram showing an example of the configuration of a conventional time-series data management system of this type. Hereinafter, a conventional time-series data management method will be described with reference to FIG.

In FIG. 14, a latest data table 1404 is shown.
Is a table for managing the latest data 1405. The time series data table 1406 is a table for managing the past time series data 1407 of the latest data table 1404.

[0004] The data updating means 1401 stores the latest data 1
In response to an instruction to insert, update, or delete 405, the latest data table updating unit 1402 updates the latest data 1405 and the time-series data table updating unit 1403 inserts time-series data 1407.

[0005] The latest data table updating means 1402 gives a time value indicating the validity period to the latest data 1405,
The latest data 1405 is updated.

The time-series data table updating means 1403 adds a time value indicating the validity period to the time-series data 1407 and stores the time-series data 1407 in the time-series data table 140.
Insert into 6.

[0007] The time-series data search means 1408 stores the time-series data set 1411 (the latest data 1405 and the time-series data 1) extracted by the time-series data extraction means 1409.
407), the time-series data search unit 1410 searches for necessary data.

The time-series data extraction means 1409 extracts the data at the time of the search target (latest data 1405 and time-series data 1407) from the latest data table 1404 and the time-series data table 1406, and extracts the time-series data set 141.
1 is generated.

[0009] The time-series data search means 1410 is a time-series data set 1 generated by the time-series data extraction means 1409.
From 411, necessary data (latest data 1405 or time-series data 1407) is searched.

[0010] As a patent gazette relating to the above-mentioned conventional technology, there is Japanese Patent Application Laid-Open No. Hei 8-287104 (name of invention: "time-series data management system").

[0011]

The above-described conventional time-series data management method has the following problems.

The first problem is that since the latest data and the time-series data are stored in separate tables, it is necessary to combine the two tables to generate a time-series data set once in data retrieval. There is a problem that extra search time is required. In particular, there is a problem that it takes a very long time to search under complicated search conditions such as when searching by combining time-series data corresponding to a plurality of latest data tables.

The second problem is that since the latest data and the time-series data are stored in different tables, it is necessary to manage the two tables of the latest data table and the time-series data table in association with each other. Also, it is necessary to maintain the consistency between these two tables in the table maintenance operation (for example, when the definition of the table changes, the definition of the latest data table and the definition of the corresponding time-series data table are simultaneously made. Need to be changed)
Therefore, there is a problem that the operation of the relational database becomes very complicated.

In view of the above, an object of the present invention is to generate a relational database capable of coexisting the latest data and past time-series data at a plurality of points in the same table, and to update the latest data. Another object of the present invention is to provide a time-series data management method that automatically saves time-series data, thereby storing and referencing data contents at a plurality of points in the past while continuing to update the relational database.

As patent publications relating to the prior art for the present invention, there are the following publications 1) and 2) in addition to the above-mentioned Japanese Patent Application Laid-Open No. 8-287104. 1) JP-A-6-35774 2) JP-A-7-325839

The technique (database management system) described in the above-mentioned JP-A-6-35774.
Manages time-series data at any point in a database file in the same database file as the database file in which the latest data is stored, thereby reducing the burden on users to manage time-series data and updating data. At the same time, by retaining time-series data, statistical processing during data update in a continuous non-stop operation environment is realized.

This prior art is similar to the present invention in that the concept of "time series data" is introduced. However, the related art is not a technology applied to a relational database, and is, of course, not a technology for centrally managing the latest data and time-series data in the same table in the relational database. Therefore, the related art differs from the time-series data management method of the present invention in its configuration, purpose, and the like.

The technique (time-series data processing device) described in the above-mentioned Japanese Patent Application Laid-Open No. 7-325839 divides a database for storing data into time-series units to form a database unit. The divided database unit is provided with means for storing data in time series units.

This prior art is similar to the present invention in that the concept of "time series data" is introduced. However, the related art stores data in a divided database unit according to a time series unit, and is not a technique for centrally managing the latest data and the time series data in the same table. Therefore, the related art differs from the time-series data management method of the present invention in its configuration, purpose, and the like.

[0020]

The time-series data management system of the present invention has a latest structure having a structure consisting of an end point and a start point for managing the validity period of data and user data arbitrarily defined by a user. A time-series table for centrally managing data and time-series data having the same structure as the latest data and holding data contents of the latest data in the past,
An entry composed of a linked key formed by the unique key and the end point and a pointer value for managing the address of the in-table data (latest data or time-series data) in the time-series table corresponding to the linked key is stored. Holds an entry composed of a unique index to be held, a link key formed by an end point and a start point, and a pointer value for managing an address of in-table data in the time series table corresponding to the link key. Validity period index, holding time management area which is an area for externally indicating a holding time as a reference for holding time series data in the time series table, and in-table data at any time in the time series table The search time management area, which is an area for externally indicating the search time indicating whether to search, and the table definition of the relational database When the generation of the time series table is instructed, the end time column and the start time column are defined in addition to the user data sequence arbitrarily defined by the user, and the time sequence table is generated. A time-series table definition for generating the validity index by defining a connection key with a time point in each entry, and adding the definition of the end time point as a part of each entry having a unique key to generate the unique index Means, and in response to an instruction to insert / update / delete the latest data in the time-series table, a holding time designated in the holding time management area at the time of inserting / updating the latest data, and "end indicating that the data is the latest data""Timevalue" is set to the value of the start time and the end time of the latest data, and time-series data is inserted into the time-series table when the latest data is updated / deleted. A time-series table updating means for performing insertion / update / deletion of the latest data according to the input / update / delete instruction and updating the unique index and the validity period index by reflecting the above processing; and In response to a search instruction, based on the search time point specified in the search time point management area, a "condition for selecting only data in the table at a specific time point" is added to the search condition to search the time-series table. When the search condition includes a unique key in the search, the value of the unique key in the unique index and the end time are evaluated, and when the search condition does not include the unique key in the search, A time-series table search unit that evaluates values of an end point and a start point in the validity period index.

Further, the time series data management system of the present invention
Latest data having a structure consisting of the end time and start time for managing the validity period of data and user data arbitrarily defined by the user, and data contents at the past time point of the latest data having the same structure as the latest data A time-series table for centrally managing time-series data holding a pointer, a link key formed by a unique key and an end point, and a pointer for managing addresses of data in the table in the time-series table corresponding to the link key. A unique index that holds an entry composed of a value, a concatenated key formed by an end time, a start time, and a non-unique key set by an arbitrary definition of the user, and the time series corresponding to the concatenated key A validity period index holding an entry composed of a pointer value for managing an address of in-table data in the table; A holding time management area, which is an area for externally instructing a holding time as a reference for holding the time series data in the time series table, and at which point in the time series table to search for in-table data. The search time point management area, which is an area for externally indicating the search time point to be indicated, and when the generation of the time series table is instructed in the table definition of the relational database, in addition to the user data string arbitrarily defined by the user, The end time column and the start time column are defined to generate the time series table, and the validity index is generated by defining a connection key of the end time, the start time, and the non-unique key in each entry, A time series table defining means for adding the definition of the end point as a part of each entry having a unique key to generate the unique index, and the latest data of the time series table. In response to the insert / update / delete instruction of the latest data, the holding time designated in the holding time management area and the “end time value indicating that it is the latest data” at the time of inserting / updating the latest data are started at the start of the latest data. The values are set to the values at the time point and the end point, time-series data is inserted into the time-series table when the latest data is updated / deleted, and insertion / update / deletion of the latest data according to the insertion / update / deletion instruction is performed. Performing a time-series table updating unit that updates the unique index and the validity period index by reflecting the above-described processing, and in response to a search instruction of the time-series table, to a search time specified in the search time management area. Based on the search condition, the time-series table is searched by adding "conditions for selecting only in-table data at a specific time" to the search condition, and the search condition includes a unique key in the search. If the search condition does not include the unique key in the search condition, the value of the unique key in the unique index and the end time are evaluated. It is also possible to adopt a configuration having time-series table search means for evaluating the values at the time point and the start time point.

More generally, the time-series data management system of the present invention uses the latest data and the time-series data having a structure having an end point and a start point which can be information for identifying the latest data and the time-series data. A time-series table for centrally managing data, a unique index for managing addresses of data in the time-series table with a concatenated key formed by a unique key and an end point for each entry, and an end for each entry. A validity period index for managing the address of the in-table data in the time-series table with a connection key formed by a time point and a start time, and when the generation of the time-series table is instructed, it is added to the user data string. The time series table is generated by defining the start time column and the end time column, and the validity period index in which the end time and the start time are defined is generated. A time-series table defining means for generating the unique index including the definition of the end time, and a value of the instruction and the value of the retention time in response to the latest data insertion / update / deletion instruction performed after the retention time is specified When updating the unique index and the validity period index by reflecting the above-mentioned processing after performing the creation / update / deletion of the latest data after adjusting the creation of the time series data based on and the insertion into the time series table In response to a series table updating means and a search instruction given after a search time is specified, considering only the search time and the end time and start time of the data in the table, "select only the data in the table at a specific time. To the time-series table using the unique index and the validity period index to search the time-series table according to the search condition. It can be said to be configured to have a time-series table search means.

[0023]

Next, the present invention will be described in detail with reference to the drawings.

(1) First Embodiment FIG. 1 is a block diagram showing a configuration of a time-series data management system according to a first embodiment of the present invention.

The time series data management method according to the present embodiment includes a time series table defining means 1, a time series table 2, a unique index 5, a validity index 6, a holding time management area 7, a time series It is configured to include a table updating means 8, a time series table searching means 12, and a search time point management area 15. The relational database is composed of the time series table 2, the unique index storage unit 5, and the validity period index 6.

The time series table 2 includes the latest data 3 having a structure composed of user data arbitrarily defined by an end time and a start time for managing the validity period of the data, and the same structure as the latest data 3. This is a table (relational database table) that unitarily manages the latest data 3 and the time-series data 4 holding the data contents at a plurality of points in the past. Note that the latest data 3 and the time-series data 4 that are centrally managed are collectively referred to as “table data”.

The unique index 5 is a unique key (a key that does not allow duplicate values such as a primary key) in the user data in the in-table data (the latest data 3 or the time-series data 4) in the time-series table 2. It holds an entry composed of a concatenated key formed from the end of the in-table data and a pointer value for managing the address of the in-table data corresponding to the concatenated key.

The validity period index 6 is a link key formed by the end point and start point of the in-table data (latest data 3 or time-series data 4) in the time-series table 2, and a table corresponding to the link key. And an entry composed of a pointer value for managing the address of the internal data.

The holding point management area 7 is an area for externally indicating a holding point (for example, a work date) as a reference for holding the time-series data 4. It should be noted that the “time point” in the present specification means a logical value set so that the value increases in a time series.

The search time point management area 15 is an area for externally designating a search time point indicating which point in the time series table 2 to search (access) in-table data.

The time series table definition means 1 enables the generation of the time series table 2 in the table definition of the relational database, and when the generation of the time series table 2 is instructed, the following 1) to
Perform the process shown in 3). 1) The end time column and the start time column are defined in addition to the user data sequence arbitrarily defined by the user, and the time series table 2 is generated (defined). 2) By defining a non-unique key connecting the end time and the start time, a validity period index 6 holding each entry having such a non-unique key is generated. 3) An end point definition is added to a unique index (index having a unique key) arbitrarily defined by the user, and a unique index 5 in which the end point definition is added as a part of each entry is generated.

The time-series table updating means 8 responds to an instruction to insert / update / delete the latest data 3 of the time-series table 2 (one of an insert instruction, an update instruction and a delete instruction) as necessary. The values of the end point and the start point of the data 3 are set, and the time series data 4 is added to the time series table 2 as necessary.
Then, the latest data 3 is inserted / updated / deleted, and the unique index 5 and the validity period index 6 defined by the time series table defining means 1 are updated by reflecting the above processing.

The time series table updating means 8 includes a valid period setting means 9, a time series data inserting means 10, and a data updating means 1.
1 is included. The time series table updating means 8 is not limited to the configuration having the three means shown in FIG. 1 as long as it performs the above-described predetermined processing.

The validity period setting means 9 inserts / updates the latest data 3 in the time series table 2 (when inserting or updating).
, The holding time is input from the holding time management area 7, the value of the holding time is set as the start time of the latest data 3 to be inserted / updated, and the maximum value (HIGH-VALUE) is set to the end time of the latest data 3. Set to.

The time-series data inserting means 10 generates the time-series table 2
When the latest data 3 is updated / deleted (when updating or deleting), the contents of the latest data 3 before update or before deletion are acquired from the time series table 2 and the start time and the holding time management area 7 are obtained. When the holding time is larger than the holding time obtained from the above, if the holding time is set at the end time of the contents of the latest data 3 before update or before deletion as the time-series data 4 Insert in Time Series Table 2.

The data updating means 11 responds to the instruction to insert / update / delete the latest data 3 in the time series table 2 by setting the values of the latest data 3 to the start time and the end time by the validity period setting means 9 and After the insertion of the time-series data 4 into the time-series table 2 by the time-series data insertion means 10 (all of these processes by the validity period setting means 9 and the time-series data insertion means 10 are not necessarily performed).
The latest data 3 is inserted / updated / deleted, and the unique index 5 and the validity period index 6 are updated by reflecting the above processing.

In response to the search instruction of the time series table 2, the time series table search means 12 adds "conditions for selecting only the data in the table at a specific time" to the search conditions, and Search for. At the time of this search, if the search condition includes the unique key in the unique index 5, the value of the unique key in the unique index 5 and the value at the end time are evaluated, and the table satisfying the search condition is evaluated. Only data is detected from the time series table 2. If the search condition does not include the unique key in the unique index 5, the values at the end point and the start point in the validity period index 6 are evaluated, and only the in-table data that satisfies the search condition is displayed in the time-series table. 2 to detect.

The time series table search means 12 includes a search condition changing means 13 and a data search means 14. Note that the time-series table search means 12 is not limited to the configuration having the two means shown in FIG. 1 as long as it performs the above-described predetermined processing.

The search condition changing means 13 inputs the search time from the search time management area 15 and sets a condition (the start time is equal to or smaller than the value at the search time and the end time is larger than the value at the search time) (a specific time). (A condition for selecting only the data in the table at the time) is added as a search condition in the search instruction. However, if the search time point is a value instructing the search for the latest data 3, the value at the search time point is evaluated as "maximum value (HIGH-VALUE)-(minus) 1".

In response to the search instruction of the time series table 2, the data search means 14 searches the time series table 2 according to the search conditions to which the conditions have been added by the search condition changing means 13, and retrieves the in-table data satisfying the search conditions. It is detected from the time series table 2.

FIG. 2 is a diagram showing a detailed configuration of the time series table 2. The time series table 2 includes an end time point column 21, a start time point column 22, and a user data string 23. The end point in the end point column 21 is data for managing the effective end point value of the in-table data (latest data 3 or time-series data 4). Further, the start time constituting the start time column 22 is data for managing the effective start time value of the data in the table. Further, the user data string 23
Are user-defined data arbitrarily defined by the user. Each in-table data in the time series table 2 has a structure having an end time in the end time column 21, a start time in the start time column 22, and user data in a user data sequence 23.

FIG. 3 is a diagram showing a detailed configuration of each entry in the unique index 5. Each entry in the unique index 5 is composed of a unique key, an end point, and a pointer value. The unique key is a key corresponding to “data that does not allow duplicate values” included in the user data in the user data string 23 in FIG. The end point is a key corresponding to the end point in the end point column 21 in FIG. The value obtained by concatenating the unique key and the end time is the unique index 5
Is the key value (the value of the concatenated key). Further, the pointer value is stored in each data in the time series table 2 (latest data 3
Alternatively, it is a value indicating the storage address of the time-series data 4).

FIG. 4 is a diagram showing a detailed configuration of each entry in the validity period index 6 in the present embodiment. Each entry in the validity period index 6 includes an end time, a start time, and a pointer value.
The end point is a key corresponding to the end point in the end point column 21 in FIG. The start time is a key corresponding to the start time in the start time column 22 in FIG.
A value obtained by connecting the end time and the start time is a key value (a value of a connected key) in the validity period index 6. Further, the pointer value is a value indicating the storage address of each in-table data (latest data 3 or time-series data 4) in the time-series table 2.

FIG. 5 is a flowchart showing the processing of the time series table defining means 1. This processing is performed in the time series table generation step 50.
1, a validity period index definition step 502, and a unique index definition step 503.

FIGS. 6 to 8 are flowcharts showing the processing of the time series table updating means 8.

The processing shown in FIG. 6 (processing based on the instruction to insert the latest data 3) includes a holding time acquisition step 601, a start time holding time setting step 602, an end time maximum value setting step 603, and a latest data time series. It comprises a table insertion step 604 and a unique index / valid period index entry registration step 605.

The processing shown in FIG. 7 (processing based on the update instruction of the latest data 3) includes a holding time acquisition step 701, a latest data acquisition step before update 702, a holding time / start time comparison judgment step 703, a time series Data creation step 704 and time series data time series table insertion step 70
5, a unique index / valid period index entry registration step 706, a latest data update step 707, and a valid period index entry update step 708.

The processing shown in FIG. 8 (processing based on an instruction to delete the latest data 3) includes a holding time acquisition step 801, a latest data before deletion step 802, a holding time / start time comparison determination step 803, a time series Data creation step 804 and time series data time series table insertion step 80
5, a unique index / valid period index entry registration step 806, a latest data deletion step 807, and a unique index / valid period index entry deletion step 808.

FIG. 9 is a flowchart showing the processing of the time series table search means 12. This processing is performed at the retrieval time point acquisition step 9
01 and search time related condition search condition addition step 902
And a search condition unique key inclusion determination step 903;
Unique key / end point evaluation step 904, in-table data search step 905, search result detection step 90
6, an end time / start time evaluation step 907, and an in-table data search / search result detection step 908.

Next, the operation of the time-series data management system according to the present embodiment configured as described above will be described in detail.

First, the operation of the time series table defining means 1 will be described.

When the generation of the time series table 2 is instructed in the table definition of the relational database shown in FIG. 1, the time series table definition means 1 performs the following processing (see FIG. 5).

First, the end time and start time for managing the validity period of the data defined by the table definition are defined by adding to the user data arbitrarily defined by the user. Then, the time series table 2 having the start time column 22 and the user data column 23 is generated (step 501).

Next, in the index definition of the time-series table 2, as a non-unique index corresponding to the end time column 21 and the start time column 22, an entry having a concatenated key connecting the end time and the start time is stored. A period index 6 is defined (step 502).

Further, the definition of the end point is added to the definition of the unique index 5 having the “unique key that does not allow duplicate values” (primary key, etc.) in the user data column 23 of the time series table 2, and the unique key and the end Unique index 5 holding an entry with a concatenated key with the time
Is defined (step 503).

Second, the operation of the time series table updating means 8 will be described.

The holding point management area 7 in FIG. 1 is an area for giving a holding point from the outside when updating the time series table 2. The possession time point is a logical value indicating a reference time point in an operation that refers to the time-series data 4 such as a statistical operation, and corresponds to, for example, a value indicating an operation date when the time-series data 4 is used in daily processing. .

When the user issues an instruction to insert / update / delete the latest data 3 together with the designation of the holding time, the time-series table updating means 8 performs the following processing 1) to 3) ( 6 to 8).

1) When the instruction to insert the latest data 3 into the time series table 2 is issued, the time series table updating means 8 performs the following processing.

The validity period setting means 9 acquires the holding time from the holding time management area 7 (step 601), and sets the holding time at the start time of the latest data 3 to be inserted (step 602).

Next, the validity period setting means 9 sets the maximum value (HIGH-VALUE) that can be taken as the end time of the in-table data at the end time of the latest data 3 to be inserted (step 603).

Thereafter, the data updating means 11 inserts the latest data 3 into the time series table 2 (step 604),
The corresponding entry is registered in the unique index 5 and the validity period index 6 (step 60).
5).

2) When an update instruction of the latest data 3 of the time series table 2 is issued, the time series table updating means 8 performs the following processing.

The validity period setting means 9 acquires the holding time from the holding time management area 7 (step 701).

At the same time, the time series data inserting means 10
Acquires the contents of the latest data 3 before updating from the time series table 2 (step 702), and compares the holding time acquired in step 701 with the start time of the latest data 3 acquired in step 702. The magnitude is compared and determined (step 703).

The time-series data inserting means 10 executes step 7
If it is determined that the value of the holding time is larger than the value of the starting time in 03, the holding time acquired by the validity period setting means 9 is set to the end time of the contents of the latest data 3 before updating. Then, the time series data 4 is created (step 704), and the time series data 4 is inserted into the time series table 2 (step 705).

At the same time, the time series data inserting means 10
Registers the entry corresponding to the time-series data 4 in the unique index 5 and the validity period index 6 (step 706).

Thereafter, the data updating means 11 sets the holding time at the start time of the latest data 3, sets the maximum value at the end time, and updates the user data of the latest data 3 (step S1). 707).

At the same time, the data updating means 11 updates the contents of the start point of the entry in the validity period index 6 having the data storage address of the latest data 3 as a pointer value at the start point of the updated latest data 3. The content is updated to have the content (step 708).

On the other hand, the data updating means 11
If it is determined in 03 that “the value at the holding time is the same as the value at the starting time”, the time-series data 4 is not inserted,
The latest data 3 is updated as it is (steps 707 and 708). If it is determined in step 703 that “the value at the possession time is smaller than the value at the start time”, it is treated as an error.

3) When an instruction to delete the latest data 3 in the time series table 2 is issued, the time series table updating means 8 performs the following processing.

The validity period setting means 9 acquires the holding time from the holding time management area 7 (step 801).

At the same time, the time series data inserting means 10
Obtains the contents of the latest data 3 before deletion from the time series table 2 (step 802), and compares the holding time acquired in step 801 with the start time of the latest data 3 acquired in step 802. The size is compared and determined (step 803).

The time-series data inserting means 10 executes step 8
If it is determined that the value of the holding time is larger than the value of the starting time in 03, the holding time acquired by the validity period setting means 9 is set to the end time of the contents of the latest data 3 before deletion. Then, the time series data 4 is created (step 804), and the time series data 4 is inserted into the time series table 2 (step 805).

At the same time, the time series data inserting means 10
Registers the entry corresponding to the time-series data 4 in the unique index 5 and the validity period index 6 (806).

Thereafter, the data updating means 11 deletes the latest data 3 (step 807).

At the same time, the data updating means 11 deletes the entry corresponding to the latest data 3 from the unique index 5 and the validity period index 6 (Step 808).

On the other hand, the data updating means 11
If it is determined in 03 that “the value at the holding time is the same as the value at the starting time”, the time-series data 4 is not inserted,
The latest data 3 is deleted as it is (steps 807 and 808). If it is determined in step 803 that “the value at the time of possession is smaller than the value at the time of start”, it is treated as an error.

Third, the operation of the time series table search means 12 will be described.

The search time management area 15 in FIG. 1 is an area for giving a search time from the outside when searching the time series table 2. The search time point is a logical value indicating the latest or past time point. Here, the "past time" is a time corresponding to the holding time set in the holding time management area 7 when the time series table 2 is updated. As the logical value indicating the latest time point, for example, an initial value relating to the time point can be considered.

When the user gives a search instruction to the time series table 2 together with the above-mentioned designation of the search time, the time series table search means 12 performs the following processing (see FIG. 9).

The search condition changing means 13 acquires the search time from the search time management area 15 (step 901), and responds to the search condition in the search instruction by saying that the start time is less than or equal to the value of the search time and the end time is (Step 902). In addition,
At this time, if the search time point is a value indicating “latest”, the search time point in the search condition is set to the “maximum value (HIGH-VAL)”.
UE) -1 ". As a result, a condition that the start time is equal to or less than “maximum value−1” and the end time is greater than “maximum value−1” is added to the search condition of the latest data 3, so that the maximum time Only the newest data 3 in which is set satisfies the search condition.

Next, the data search means 14 determines whether or not the search condition includes a unique key (step 903).

If it is determined in step 903 that “the search condition includes a unique key”, the data search means 14 searches the unique index 5 and finds the unique key and the unique key in each entry in the unique index 5. The value at the end point is evaluated (step 904).

As a result of the evaluation in step 904, the data search means 14 searches the time series table 2 for the entry of the unique index 5 whose unique key and end value satisfy the search condition based on the pointer value in the entry. In-table data (latest data 3 or time-series data 4) to be searched is searched (step 905), and only in-table data whose start time satisfies the search condition is selected from the in-table data detected in the search, and the in-table data is selected. The data is detected as a search result for the search instruction (step 906).

The data search means 14 determines in step 9
When it is determined that the search condition does not include the unique key in 03, the validity period index 6 is searched,
The values of the end point and the start point in each entry in the validity period index 6 are evaluated (step 907).

As a result of the evaluation in step 907, the data search means 14 determines, for the entry of the validity period index 6 whose end time and start time satisfy the search condition, from the time series table 2 based on the pointer value in the entry. The corresponding in-table data (latest data 3 or time-series data 4) is detected as a search result for the search instruction (step 9).
08).

(2) Second Embodiment FIG. 1 is also a block diagram showing a configuration of a time-series data management system according to a second embodiment of the present invention.

The time-series data management system according to this embodiment has the same basic configuration as that of the first embodiment, but the configuration of each entry in the validity period index 6 is as shown in FIG. The third embodiment differs from the first embodiment in the point shown in FIG.

FIG. 10 is a diagram showing a detailed configuration of each entry in the validity period index 6 in the present embodiment. Each entry in the validity period index 6 according to the present embodiment includes a non-unique key, an end time, a start time, and a pointer value. The non-unique key is a key corresponding to “data that allows duplicate values” included in the user data in the user data string 23 in FIG. 2, and is arbitrarily defined by the user. Also,
The end point is a key corresponding to the end point in the end point column 21 in FIG. Further, the start time is a key corresponding to the start time in the start time column 22 in FIG. The value obtained by concatenating the non-unique key, the end point, and the start point is the key value (concatenated key value) in the validity period index 6. In addition, the pointer value is a value indicating the storage address of each in-table data (latest data 3 or time-series data 4) in the time-series table 2.

FIG. 11 is a flowchart showing the processing of the time series table defining means 1 in the present embodiment. This processing includes a time series table generation step 1101, a validity period index definition step 1102, and a unique index definition step 1103.

FIG. 12 is a flowchart showing the processing of the time series table search means 12 in the present embodiment. This process
Search time acquisition step 1201, search time related condition search condition addition step 1202, search condition unique key inclusion determination step 1203, unique key / end time evaluation step 1204, in-table data search step 12
05, a search result detection step 1206, a non-unique key / end time / start time evaluation step 1207, and a table data search / search result detection step 1208.

FIGS. 6 to 8 are also flow charts showing the processing of the time series table updating means 8 in the present embodiment.

Next, the operation of the time-series data management system according to the present embodiment configured as described above will be described in detail.

First, the operation of the time series table defining means 1 will be described.

When the generation of the time series table 2 is instructed in the table definition of the relational database shown in FIG. 1, the time series table definition means 1 performs the following processing (see FIG. 11).

First, the end time and start time for managing the validity period of the data defined by the table definition are defined by adding to the user data arbitrarily defined by the user. Then, a time series table 2 having a start time column 22 and a user data column 23 is generated (step 1101).

Next, in the index definition of the time series table 2, the definition of the end time and the start time is added to the non-unique index (index having the non-unique key) defined by the user, and the non-unique key is added. A validity period index 6 for holding an entry having a concatenated key connecting the end point and the start point is defined (step 1102).

Further, the definition of the end point is added to the definition of the unique index 5 having the “unique key whose value is not allowed to be duplicated” (primary key or the like) in the user data column 23 of the time series table 2, and the unique key and the end Unique index 5 holding an entry with a concatenated key with the time
Is defined (step 1103).

Second, the operation of the time series table updating means 8 will be described.

The operation of the time series table updating means 8 is the same as the operation of the time series table updating means 8 in the first embodiment (see FIGS. 6 to 8).

Third, the operation of the time series table search means 12 will be described.

As described in the description of the first embodiment, the search time point management area 15 in FIG. 1 is an area for externally giving a search time point when searching the time series table 2. . The search time point is a logical value indicating the latest or past time point. Here, the "past time" is a time corresponding to the holding time set in the holding time management area 7 when the time series table 2 is updated. As the logical value indicating the latest time point, for example, an initial value relating to the time point can be considered.

When the user gives a search instruction to the time series table 2 together with the above-mentioned designation of the search time, the time series table search means 12 performs the following processing (see FIG. 12).

The search condition changing means 13 acquires the search time from the search time management area 15 (step 1201), and responds to the search condition in the search instruction by saying that “the start time is less than or equal to the value of the search time and the end time is Is greater than the value at the time of retrieval "(step 1202). At this time, if the search time point is a value indicating “latest”, the search time point in the search condition is set to “maximum value (HIGH−
VALUE) -1 ".

Next, the data search means 14 determines whether or not the search condition includes a unique key (step 1203).

The data search means 14 determines in step 1203
If it is determined that the search condition includes a unique key, the unique index 5 is searched, and the unique key in each entry in the unique index 5 and the value at the end point are evaluated (step 1204). .

The data search means 14 determines in step 1204
As a result of the evaluation, for the entry of the unique index 5 whose unique key and the value at the end time satisfy the search condition, the corresponding in-table data (latest data 3 or time series data) from the time series table 2 based on the pointer value in the entry. Data 4) is searched (step 1205), and only the in-table data whose start point satisfies the search condition is selected from the in-table data detected in the search, and the in-table data is detected as a search result in response to the search instruction (step 1205). Step 1206).

Further, the data search means 14 executes step 1
In 203, "the search condition does not include the unique key"
If it is determined, the validity period index 6 is searched and the non-unique key in each entry in the validity period index 6 and the values at the end time and the start time are evaluated (step 1207).

The data search means 14 determines in step 1207
As a result of the evaluation, for the entry of the validity period index 6 in which the values of the non-unique key and the end time and the start time satisfy the search condition, the corresponding in-table data (latest data) from the time series table 2 based on the pointer value in the entry. Data 3
Alternatively, the time series data 4) is detected as a search result for the search instruction (step 1208).

As described above, in the time-series data management method according to the present embodiment, instead of the validity period index 6 in which the concatenated key is constituted only by the end point and the start point,
By adding the end time and the start time to the non-unique key arbitrarily defined by the user, the validity period index 6 that forms the concatenated key is generated. Thus, in the search using the validity period index 6, it is possible to efficiently narrow down the data in the table to be searched by both the non-unique key and the validity period (end time and start time).

(3) Third Embodiment FIG. 13 is a block diagram showing a configuration of a time-series data management system according to a fourth embodiment of the present invention. The time-series data management method according to the present embodiment includes a recording medium 100 in which a predetermined program (time-series data management program) is recorded in addition to the time-series data management method according to the above-described first embodiment. The points are different. This recording medium 1
00 may be a magnetic disk, a semiconductor memory, or another recording medium.

The time series data management program is read from the recording medium 100 into the computer 200, and is stored in the relational database (time series table 2, unique index 5, and validity period index 6), holding time management area 7, and search time management area 15. The operation of the computer 200 is assumed to be based on the existence of
Control is performed as the time series table updating means 8 and the time series table search means 12. The operation of the computer 200 under the control of the time series data management program is exactly the same as the operation of the time series table definition means 1, the time series table update means 8, and the time series table search means 12 in the first embodiment. , The detailed explanation is omitted.

(4) Fourth Embodiment FIG. 13 is also a block diagram showing a configuration of a time-series data management system according to a fourth embodiment of the present invention. The time-series data management method according to the present embodiment includes a recording medium 100 in which a predetermined program (time-series data management program) is recorded in addition to the time-series data management method according to the second embodiment. The points are different. The recording medium 100 may be a magnetic disk, a semiconductor memory, or another recording medium.

The time-series data management program is read from the recording medium 100 into the computer 200, and is stored in a relational database (time-series table 2, unique index 5, and validity period index 6), holding time management area 7, and search time management area 15. The operation of the computer 200 is assumed to be based on the existence of
Control is performed as the time series table updating means 8 and the time series table search means 12. The operation of the computer 200 under the control of the time-series data management program is exactly the same as the operation of the time-series table definition means 1, the time-series table update means 8, and the time-series table search means 12 in the second embodiment. , The detailed explanation is omitted.

[0116]

As described above, the present invention has the following effects.

First, by adding end time and start time to tables and indexes of the relational database, etc., the latest data and past time-series data at a plurality of time points are unitarily managed in the same table (time-series table). It is possible to use time-series data with the same user interface as the latest data,
The management and operation of time-series data can be simplified (for example, when the latest data is updated, the data contents before the update are inserted into the time-series table as time-series data. Data can be stored).

It should be noted that supplementing the above, the latest data and a plurality of time-series data corresponding to the latest data can be stored (unified management) in the same table (time-series table) as follows. Based on 1) and 2). 1) The unique key in the unique index can ensure uniqueness between certain latest data and other latest data. 2) By retaining the end point as a part of the concatenated key in the unique index, only one latest data or time series data with the same end point exists for one latest data. Data values do not overlap with time-series data at a plurality of past time points, and uniqueness between the data can be ensured.

Second, a time-series table defining means for automatically performing a definition necessary for managing the time-series data, and a method for automatically creating time-series data at the time of updating the latest data based on the assignment at the time of holding. By providing the sequence table updating means and the time series table search means for automatically searching the data in the table at the required time based on the assignment of the search time, the user can add no special control to the application program or the like. Thus, there is an effect that time series data can be easily collected and used.

Third, by retaining the time-series data at the time of updating the latest data, the update operation of updating the latest data of the relational database and the statistical operation of referring to the time-series data of the past time can be performed in parallel. (It is possible to save and refer to data contents at a plurality of points in the past while continuing to update the database), so that it is not necessary to stop the online information processing system for statistical work ( 24 hours operation of the online information processing system becomes possible).

By providing a validity period index and performing a search for the latest data or time-series data in the time-series table via the validity period index, access to data at an unnecessary point in the time-series table does not occur. For this reason, it is possible to prevent performance degradation (decrease in data search speed) due to centralized management of the latest data and time-series data (this is more remarkable when a non-unique key is provided in the validity period index).

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a time-series data management system according to a first embodiment and a second embodiment of the present invention.

FIG. 2 is a diagram showing a detailed configuration of a time series table in FIG.

FIG. 3 is a diagram showing a detailed configuration of each entry in a unique index in FIG. 1;

FIG. 4 is a diagram showing a detailed configuration of each entry in a validity period index according to the first embodiment of the present invention.

FIG. 5 is a flowchart showing processing of a time-series table definition unit according to the first embodiment of the present invention.

FIG. 6 is a flowchart showing processing of a time-series table updating unit according to the first embodiment and the second embodiment of the present invention.

FIG. 7 is a flowchart showing processing of a time-series table updating unit in the first embodiment and the second embodiment of the present invention.

FIG. 8 is a flowchart showing processing of a time-series table updating unit in the first embodiment and the second embodiment of the present invention.

FIG. 9 is a flowchart showing processing of a time-series table search unit according to the first embodiment of the present invention.

FIG. 10 is a diagram showing a detailed configuration of each entry in a validity period index according to the second embodiment of the present invention.

FIG. 11 is a flowchart showing processing of a time-series table definition unit according to the second embodiment of the present invention.

FIG. 12 is a flowchart showing processing of a time-series table search unit according to the second embodiment of the present invention.

FIG. 13 is a block diagram illustrating a configuration of a time-series data management system according to a third embodiment and a fourth embodiment of the present invention.

FIG. 14 is a block diagram showing a configuration of an example of a conventional time-series data management method.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Time series definition means 2 Time series table 3 Latest data 4 Time series data 5 Unique index 6 Validity period index 7 Holding time management area 8 Time series table updating means 9 Validity period setting means 10 Time series data inserting means 11 Data updating means 12 Time series table search means 13 Search condition changing means 14 Data search means 15 Search time management area 21 End time string 22 Start time string 23 User data string 100 Recording medium 200 Computer

Claims (9)

[Claims] 1. A time series table for centrally managing latest data and time series data having a structure having an end time and a start time which can be information for identifying the latest data and time series data, and a unique key for each entry. A unique index for managing an address of data in the table in the time-series table with a connection key formed by an end time, and a connection key formed by an end time and a start time for each entry. The validity period index for managing the address of the data in the table in the series table, and when the generation of the time series table is instructed, a start time column and an end time column are defined in addition to the user data column to define the time series table. And generate
A time-series table defining means for generating the validity index in which the end time and the start time are defined, and generating the unique index including the definition of the end time, and inserting the latest data performed when the holding time is specified In response to the / update / delete instruction, the latest data is inserted / updated / deleted after adjusting the creation of time-series data based on the instruction and the value at the time of holding and the insertion into the time-series table. A time-series table updating unit that updates the unique index and the validity period index by reflecting
In response to a search instruction given after the search time has been specified,
In consideration of the search time and the end time and start time of the in-table data, “condition for selecting only in-table data at a specific time” is added to the search condition, and the unique index and the validity period index are used. And a time-series table search means for searching the time-series table according to the search condition. 2. The latest data having a structure composed of an end time and a start time for managing the validity period of data and user data arbitrarily defined by a user, and the latest data having the same structure as the latest data and having the same structure as the latest data. A time series table for centrally managing the time series data holding the data contents at the point of time, a linked key formed by a unique key and an end point, and the in-table data in the time series table corresponding to the linked key. A unique index that holds an entry composed of a pointer value for managing an address, a concatenated key formed by an end point and a start point, and an address of in-table data in the time-series table corresponding to the concatenated key. A validity period index for holding entries composed of pointer values to be managed, and a reference for holding time-series data in the time-series table The holding time management area, which is an area for externally instructing the holding time, and an area for externally instructing a search time indicating which point in the time series table to search for in-table data. When the generation of the time series table is instructed in the search time management area and the table definition of the relational database, the end time column and the start time column are defined in addition to the user data columns arbitrarily defined by the user, and Generate a sequence table, generate the validity period index by defining a concatenated key between the end point and the start point in each entry, and add the definition of the end point as a part of each entry having a unique key. A time series table defining means for generating the unique index, and inserting / updating the latest data in response to an instruction to insert / update / delete the latest data in the time series table. The retention time specified in the retention time management area and the “end value indicating the latest data” are set to the start time and the end time of the latest data, and when the latest data is updated / deleted, The time series data is inserted into the time series table, the latest data is inserted / updated / deleted according to the insertion / update / deletion instruction, and the unique index and the validity period index are reflected by reflecting the above processing. A time-series table updating means for updating, and, in response to a search instruction of the time-series table, “conditions for selecting only in-table data at a specific time” based on a search time specified in the search time management area Is added to the search condition and the time-series table is searched. If the search condition includes a unique key in the search, the unique key in the unique index is A time-series table search unit that evaluates a value at the end point and evaluates values at the end point and the start point in the validity period index when the search condition does not include a unique key in the search. A time-series data management method characterized by the following. 3. When the latest data of the time series table is inserted / updated, a holding time is input from the holding time management area, and the value of the holding time is set as a start time of the latest data to be inserted / updated. Validity period setting means in the time series table updating means for setting a value to the end point of the latest data, and updating / deleting the latest data of the time series table, the latest data before updating / deleting from the time series table. And comparing the start time of the latest data with the holding time obtained from the holding time management area. If the holding time is larger in the comparison, the end of the latest data before update / deletion is obtained. A time-series data inserting unit in the time-series table updating unit for inserting the holding time set at the time as the time-series data into the time-series table, and responding to an instruction to insert / update / delete the latest data in the time-series table. hand After the setting of the values at the start and end points of the latest data by the validity period setting means and the insertion of the time series data into the time series table by the time series data insertion means, the insertion / update / update of the latest data is performed. Data is deleted in the time series table updating means for updating the unique index and the validity period index by reflecting the above processing, and a search time is input from the search time management area, and the search time is If the value is not a value instructing the search for the latest data, a condition that “the start time is equal to or less than the value at the search time and the end time is greater than the value at the search time” is added to the search condition, and the search time is If the value instructs the search, the value at the search time is evaluated as “maximum value−1”, and then “the start time is less than or equal to the value at the search time and the end time is A search condition changing means in the time series table search means for adding a condition "larger than the value at the time of the search" to the search conditions, and a condition is added by the search condition changing means to the search instruction of the time series table. 3. A data search means in a time series table search means for searching the time series table according to the search condition, and detecting in-table data satisfying the search condition from the time series table. Time series data management method. 4. The latest data having a structure composed of an end time and a start time for managing the validity period of data and user data arbitrarily defined by a user, and the latest data having the same structure as the latest data and having the same structure as the latest data. A time series table for centrally managing the time series data holding the data contents at the point of time, a linked key formed by a unique key and an end point, and the in-table data in the time series table corresponding to the linked key. A unique index that holds an entry consisting of a pointer value that manages an address, a concatenated key formed by an end time, a start time, and a non-unique key set by a user's arbitrary definition and its concatenated key A validity period holding an entry composed of a pointer value for managing the address of the in-table data in the corresponding time-series table. Index, a holding time management area which is an area for externally indicating a holding time as a reference for holding the time series data in the time series table, and searching data in the table at any time in the time series table. A search time point management area, which is an area for externally instructing a search time point indicating whether or not to perform, and when generation of the time series table is instructed in the table definition of the relational database, a user data string arbitrarily defined by a user In addition to the end time column and the start time column, the time series table is generated by defining the end time column and the start time column, and the validity index is defined by defining a connection key of the end time, the start time, and the non-unique key in each entry. A time-series table defining means for generating and adding the end point definition as part of each entry having a unique key to generate the unique index; In response to the latest data insertion / update / deletion instruction, the retention time specified in the retention time management area and the “end value indicating the latest data” at the time of insertion / update of the latest data are updated to the latest data. The values are set at the start time and end time of the data, time-series data is inserted into the time-series table when the latest data is updated / deleted, and the latest data is inserted / updated according to the insertion / update / deletion instruction. And a time-series table updating means for performing / deleting and updating the unique index and the validity period index by reflecting the above-mentioned processing, and a time-series table search instruction specified in the search time management area in response to the time-series table search instruction. Based on the search time point, “conditions for selecting only the data in the table at a specific time point” is added to the search conditions to search the time-series table. If a unique key is included, the value of the unique key in the unique index and the value of the end time are evaluated. A time-series data management method, comprising: a time-series table search means for evaluating values of a key, an end point, and a start point. 5. When the latest data of the time series table is inserted / updated, a holding time is input from the holding time management area, and the value of the holding time is set as a start time of the latest data to be inserted / updated. Validity period setting means in the time series table updating means for setting a value to the end point of the latest data, and updating / deleting the latest data of the time series table, the latest data before updating / deleting from the time series table. And comparing the start time of the latest data with the holding time obtained from the holding time management area. If the holding time is larger in the comparison, the end of the latest data before update / deletion is obtained. A time-series data inserting unit in the time-series table updating unit for inserting the holding time set at the time as the time-series data into the time-series table, and responding to an instruction to insert / update / delete the latest data in the time-series table. hand After the setting of the values at the start and end points of the latest data by the validity period setting means and the insertion of the time series data into the time series table by the time series data insertion means, the insertion / update / update of the latest data is performed. Data is deleted in the time series table updating means for updating the unique index and the validity period index by reflecting the above processing, and a search time is input from the search time management area, and the search time is If the value is not a value instructing the search for the latest data, a condition that “the start time is equal to or less than the value at the search time and the end time is greater than the value at the search time” is added to the search condition, and the search time is If the value instructs the search, the value at the search time is evaluated as “maximum value−1”, and then “the start time is less than or equal to the value at the search time and the end time is A search condition changing means in the time series table search means for adding a condition "larger than the value at the time of the search" to the search conditions, and a condition is added by the search condition changing means to the search instruction of the time series table. 5. A data search means in a time series table search means for searching the time series table according to the search condition and detecting in-table data satisfying the search condition from the time series table. Time series data management method. 6. The latest data having a structure composed of an end time and a start time for managing the validity period of data and user data arbitrarily defined by a user, and the latest data having the same structure as the latest data and having the same structure as the latest data. A time series table for centrally managing the time series data holding the data contents at the time point, a connection key formed by the unique key and the end time, and the address of the data in the table in the time series table corresponding to the connection key A unique index that holds an entry composed of a pointer value that manages a link key, a link key formed by an end point and a start point, and an address of data in the time series table corresponding to the link key. Validity index holding an entry composed of a pointer value and a reference for holding time-series data in the time-series table Retention time management area, which is an area for externally instructing the retention time, and search time, which is an area for externally instructing the retrieval time indicating which point in the time series table to retrieve in-table data, In an information processing system having a management area, a computer is
When the generation of the time series table is instructed in the table definition of the relational database, the end time column and the start time column are defined in addition to the user data columns arbitrarily defined by the user, and the time series table is generated. Generating the validity index by defining a concatenated key between the end point and the start point in each entry, and adding the end point definition as part of each entry having a unique key to generate the unique index The time-series table defining means performs the insertion / update / deletion instruction of the latest data in the time-series table, and the holding time specified in the holding time management area at the time of the insertion / update of the latest data and “the latest data. At the end of the latest data ”is set to the values at the start and end of the latest data. Time series data is inserted, the latest data is inserted / updated / deleted according to the insertion / update / deletion instruction, and the unique index and the validity period index are updated by reflecting the above processing. In response to the table update means and the search instruction of the time-series table, based on the search time point specified in the search time point management area, a "condition for selecting only the data in the table at a specific time point" is set as a search condition. Add and search the time series table,
When the search condition includes the unique key in the search condition, the value of the unique key in the unique index and the end time are evaluated. When the search condition does not include the unique key in the search, A recording medium that records a program for functioning as a time-series table search unit that evaluates values of an end point and a start point in a validity period index. 7. A computer inputs a holding time from the holding time management area when inserting / updating the latest data of the time series table, and sets a value of the holding time as a start time of the latest data to be inserted / updated. The validity period setting means in the time series table updating means for setting the maximum value to the end point of the latest data, when updating / deleting the latest data of the time series table, The content of the latest data is obtained, and the start time of the latest data is compared with the holding time obtained from the holding time management area. If the holding time is larger in the comparison, the latest data before update / deletion is obtained. Time-series data inserting means in the time-series table updating means for inserting the holding time at the end point of the time series data as time-series data into the time-series table updating means, and inserting / updating / deleting the latest data in the time-series table In response to the instruction, after setting the values to the start time and end time of the latest data by the validity period setting means and inserting time-series data into the time-series table by the time-series data insertion means, Insert /
Data is updated / deleted and the unique index and the validity period index are updated by reflecting the above processing. The data updating means in the time series table updating means, and a search time point is inputted from the search time point management area, and the search time point is inputted. Is not a value instructing the search for the latest data, a condition that “the start time is equal to or less than the value at the search time and the end time is greater than the value at the search time” is added to the search condition, and the search time is If the value indicates a data search, the value at the search time is evaluated as “maximum value−1”, and then “the start time is equal to or less than the value at the search time and the end time is greater than the value at the search time” "Is added to the search condition by the search condition changing means in the time series table search means, and the search condition changing means adds the condition to the search instruction of the time series table. A program for causing the time-series table to be searched according to the searched search condition and for functioning as a data search unit in a time-series table search unit that detects in-table data satisfying the search condition from the time-series table. 6
The recording medium according to the above. 8. The latest data having a structure composed of an end time and a start time for managing a validity period of data and user data arbitrarily defined by a user, and the latest data having the same structure as the latest data and having the same structure as the latest data. A time series table for centrally managing the time series data holding the data contents at the time point, a connection key formed by the unique key and the end time, and the address of the data in the table in the time series table corresponding to the connection key Unique index that holds an entry consisting of a pointer value that manages a link key, a link key formed by an end time, a start time, and a non-unique key set by an arbitrary definition of a user, and a key corresponding to the link key Validity period index holding an entry consisting of a pointer value for managing the address of the data in the table in the time series table. Search for a holding time management area, which is an area for externally indicating a holding time as a reference for holding the time series data in the time series table, and data in the table at any time in the time series table. In an information processing system having a search time point management area that is an area for externally indicating a search time point indicating whether a computer is instructed to generate the time series table in the table definition of the relational database, The time series table is generated by defining an end time column and a start time column in addition to an arbitrarily defined user data column, and a connection key of the end time, the start time, and the non-unique key is defined in each entry. Thus, the validity period index is generated, and a definition of an end point is added as a part of each entry having a unique key, thereby defining the unique index. Sequence table definition means when generating a box for insert / update / delete command of the most recent data of the time series table of the latest data insertion /
At the time of update, the holding time specified in the holding time management area and the `` end time value indicating the latest data '' are set to the start time and end time values of the latest data,
When the latest data is updated / deleted, time-series data is inserted into the time-series table, and the latest data is inserted / updated / deleted in accordance with the insertion / update / deletion instruction. The time-series table updating means for updating the unique index and the validity period index, and in response to a search instruction of the time-series table, "data in the table at a specific time point" based on a search time point specified in the search time point management area. Condition for selecting only) is added to the search condition and the time-series table is searched.If the search condition includes a unique key in the search, the unique key in the unique index, the end time, Evaluates the value of
If a unique key is not included in the search condition at the time of the search, a program for functioning as a time-series table search means for evaluating the values of the non-unique key in the validity period index, the end time, and the start time is recorded. recoding media. 9. The computer inputs a holding time from the holding time management area when inserting / updating the latest data of the time series table, and sets a value of the holding time as a start time of the latest data to be inserted / updated. The validity period setting means in the time series table updating means for setting the maximum value to the end point of the latest data, when updating / deleting the latest data of the time series table, The content of the latest data is obtained, and the start time of the latest data is compared with the holding time obtained from the holding time management area. If the holding time is larger in the comparison, the latest data before update / deletion is obtained. Time-series data inserting means in the time-series table updating means for inserting the holding time at the end point of the time series data as time-series data into the time-series table updating means, and inserting / updating / deleting the latest data in the time-series table In response to the instruction, after setting the values to the start time and end time of the latest data by the validity period setting means and inserting time-series data into the time-series table by the time-series data insertion means, Insert /
Data is updated / deleted and the unique index and the validity period index are updated by reflecting the above processing. The data updating means in the time series table updating means, and a search time point is inputted from the search time point management area, and the search time point is inputted. Is not a value instructing the search for the latest data, a condition that “the start time is equal to or less than the value at the search time and the end time is greater than the value at the search time” is added to the search condition, and the search time is If the value indicates a data search, the value at the search time is evaluated as “maximum value−1”, and then “the start time is equal to or less than the value at the search time and the end time is greater than the value at the search time” "Is added to the search condition by the search condition changing means in the time series table search means, and the search condition changing means adds the condition to the search instruction of the time series table. A program for causing the time-series table to be searched according to the searched search condition and for functioning as a data search unit in a time-series table search unit that detects in-table data satisfying the search condition from the time-series table. 8
The recording medium according to the above.