Database Management Program

TECHNICAL FIELD

The present invention relates to a database management system (DBMS) in which data space is segmentalized, in particular to a database management program of a DBMS for an embedded device.

BACKGROUND ART

With regard to an embedded device having large-capacity storage, list fetch is used frequently in many applications such as a car navigation system, a music player and an HDD recorder.

Especially in the car navigation system, a dedicated file system in which emphasis is put to performing a real-time processing of list fetch using an I/O device performing a seek processing extremely slowly such as CD/DVD is used, and contrivance is made so that data to be list-displayed is positioned in sequential space on the disk. However, this dedicated file system is designed for a device performing no update operation, and in a case where update is performed using a device capable of update operation represented by an HDD, all the data are rearranged at each update operation.

Further, in the DBMS, as a function to allocate data in sequential space, there is a clustered table that guarantees that data is sorted physically with respect to a specified clustered key column structuring a table. This is shown in Non-Patent Document 1, for example.

Here, an example of a conventional embodiment of the clustered table in a DBMS in which storage space management composed of a page and a segment shown in FIG. 4 is assumed is explained hereinafter.

Here, a page (5002) is a physical unit representing a minimum unit of data I/O to storage space (5001) and each of segments (5003, 5004) is a physical unit representing a minimum unit in allocation of space to a table and an index. The segment is composed of a plurality of the pages. Relation between the clustered table and the storage space is shown in FIG. 5. The clustered table (6001) is a clustered table in which a sequential number column (6002) is set to a clustered key column and segments on storage space (6007) are allocated. Columns (6003, 6004, 6005) of the clustered table (6001) are stored in a segment (6014) on the storage space (6007), and another segment (6015) is allocated to a column (6006). Thus, in a case where the sequential number column is set to a clustered key column, data is sorted physically with respect to the sequential number column. In segments (6014, 6015), empty space (6012, 6013) are secured for future insert operation, respectively.

When insert operation to the clustered table is performed, since it is required to maintain a state in which data is sorted physically, insertion is performed by a method different from that for a normal table. The method of performing insert operation to the clustered table is explained with reference to a flow chart shown in FIG. 6. In an insert processing (7001) performing the insert operation, firstly, size of a row to be inserted and size of empty space of storage space are compared and judgment (7002) whether the insertion can be performed or not is made. Here, if the insertion cannot be performed, error judgment is made (7012) and the insert processing is ended (7009). If the insertion can be performed, a segment including a largest row among rows having a clustered key which is equal or smaller is searched in the storage space (7003) and judgment (7004) whether there is empty space to insert a row in the segment or not is made. If there is the empty space, a procedure of A (7013) is carried out. Here, firstly, a lock of an index related to a corresponding table is gotten (7005), and then, the row is inserted into the empty space in the segment (7006), the index is updated (7007), the lock of the index is released (7008) and the insert processing is ended (7009). An operation example in a case where a row having a clustered key of 12 is inserted according to the procedure of A (7013) is shown in FIG. 7. Here, since there is still empty space (8005) in the segment (8002) including a row (8004) having the same clustered key, the row (8006) having the clustered key of 12 is inserted into the empty space in the same segment.

On the other hand, if there is no empty space in the judgment (7004) whether there is empty space to insert a row in the same segment or not, a procedure of B (7014) is carried out. Here, firstly, the lock of the index related to a corresponding table is gotten (7010), and then, the row is inserted into a new segment (7011), the lock of the index is released (7007) and the insert processing is ended (7009). An operation example in a case where a row having a clustered key of 12 is inserted according to the procedure of B (8014) is shown in FIG. 8. Here, since there is no empty space in the segment (9002) including a row (9004) having the same clustered key, a new segment (9005) is secured and the row (9006) having the clustered key of 12 is inserted.

By use of the clustered table, in performing a search with a range of the clustered key value designated, since storage I/O is sequentialized, the search is sped up. However, on the other hand, a state in which data is sorted physically is broken by the update and the insert operation over empty space at preparation of the table, and therefore, periodical maintenance is required. This maintenance is also rearrangement of all data in the same manner as in the update operation in the dedicated file system described above.

In the DBMS for embedded application, in order to achieve data I/O performance comparable to that of the dedicated file system incapable of update of the conventional system, it is necessary to arrange data in physical sequential space in the same manner. However, in the embedded application, since maintenance-free and long-term performance stability is important, an existing clustered table cannot be used as it is. Therefore, a table structure which is specialized for data fetch of a pattern called list fetch in the embedded application and improving I/O performance is desired.

Here, a characteristic of the list fetch is explained taking car navigation application as an example, with reference to FIG. 9. In the car navigation system, intersection data is handled as one of map information. The intersection data is divided into sections (10001) each having a unique identifier and managed. In the car navigation application, list fetch of the intersection data is required to have real-time property most strongly. This is performed at a route search, and there are two patterns described below.

1. list fetch of intersections in a section (10002) including own car

2. list fetch of intersections in 9 adjacent sections (10003) around own car

Here, a pattern in which all data is arranged in sequential space in a case where a file system dedicated for car navigation is used is the pattern of 1, and it is impossible to arrange all data including the pattern of 2 in the sequential space. Therefore, also in a DBMS specialized for embedded application, it is enough if it is guaranteed that the pattern of 1 is arranged in the sequential space.

And, as a characteristic common to the two patterns, it is pointed out that a search condition is an equal condition of a section number and that a plurality of intersections having the same section number exists in search target data. By putting this characteristic otherwise in terms of DBMS, it becomes a search of an equal condition with respect to columns including a duplicated key.

Note that, a similar list fetch appears frequently also in application other than the car navigation application. For example, list fetch such as “list of music titles by an identical artist” and “list of music titles contained in an identical album” in music player application and “list of programs of * month * day” and “list of programs of a channel *” in HDD recorder application appear.

Non-Patent Document 1: [Product Manual] Scalable Database Server HiRDB Version 7 System Introduction and Design Guide (For UNIX(R)) 3000-6-272, 12.9 Specification of Clustered Key”, pp. 342-343

DISCLOSURE OF THE INVENTION

However, in a case of managing a single or plurality of sets of rows having values of the column A duplicated using relational database, even if a row having a value of the column A duplicated is arranged in sequential space as an initial state in order to speed-up a search having a condition of the column A, by repeating insert, deletion and update of rows, a state in which the row having the value of the column A duplicated is arranged in the sequential space is broken and time required for the search having the condition of the column A is lengthened, and therefore, in a case of solving this problem by a conventional method, rearrangement of all rows is required and long time is taken.

Accordingly, an object of the present invention is to provide a database management program in which the time required for the search having the condition of the column A is not lengthened even if insert, deletion and update of rows are repeated even in the case where the row having the value of the column A duplicated is arranged in the sequential space as an initial state in order to speed up the search having the condition of the column A, and the rearrangement of all rows taking long time in the case of solving this problem by the conventional method is not required.

In the database management program according to the present invention, when no empty sequential space is found in space where the rows having the same value of the column A is arranged on storage in a case of performing insertion, deletion and update, only a set of rows having the value of the column A duplicated is arranged in the sequential space using algorithm having a characteristic of including a procedure to search for sequential empty space having capacity larger than the whole rows having the same value of the column A, and a procedure to move the whole rows having the same value of the column A to the empty space in a case where the empty space is found by the search.

According to the present invention, when managing a single or plurality of sets of rows having values of the column A duplicated using the relational database, in a case where a row having a value of the column A duplicated is arranged in sequential space as an initial state in order to speed-up the search having the condition of the column A, even if insertion, deletion and update of a plurality of rows are repeated, the time required for the search having the condition of the column A is not lengthened, and the rearrangement of all rows taking long time in the case of solving this problem by the conventional method is not required.

BRIEF DESCRIPTIONS OF THE DRAWINGS

FIG. 1 is a block diagram showing a structure of a database management system according to a first embodiment of the present invention;

FIG. 2 is a data configuration diagram showing a duplicated ID sequential arrangement table identification table according to the first embodiment of the present invention;

FIG. 3 is a configuration diagram showing a duplicated ID sequential arrangement table ID management table according to the first embodiment of the present invention;

FIG. 4 is an explanatory diagram showing a storage space management method composed of a page and a segment in an embodiment of a clustered table of a conventional system;

FIG. 5 is an explanatory diagram showing relation between the clustered table and storage space in the embodiment of the clustered table of the conventional system;

FIG. 6 is an explanatory diagram showing a flow chart of insert operation to the clustered table in the embodiment of the clustered table of the conventional system;

FIG. 7 is an explanatory diagram showing an example of a procedure A in the flow chart of the insert operation to the clustered table in the embodiment of the clustered table of the conventional system;

FIG. 8 is an explanatory diagram showing an example of a procedure B in the flow chart of the insert operation to the clustered table in the embodiment of the clustered table of the conventional system;

FIG. 9 is an explanatory diagram showing a characteristic of list fetch of the conventional system taking car navigation application as an example;

FIG. 10 is an explanatory diagram showing an example of a structure of a table including a duplicated key according to the first embodiment of the present invention;

FIG. 11 is an explanatory diagram showing a structure of a duplicated ID sequential arrangement table on storage space according to the first embodiment of the present invention;

FIG. 12 is an explanatory diagram showing an example of a schematic view of the structure of the duplicated ID sequential arrangement table on the storage space according to the first embodiment of the present invention;

FIG. 13 is an explanatory diagram showing a flow chart of insert operation to the duplicated ID sequential arrangement table according to the first embodiment of the present invention;

FIG. 14 is an explanatory diagram showing an example of a procedure A in the flow chart of the insert operation to the duplicated ID sequential arrangement table according to the first embodiment of the present invention;

FIG. 15 is an explanatory diagram showing an example of a procedure B in the flow chart of the insert operation to the duplicated ID sequential arrangement table according to the first embodiment of the present invention;

FIG. 16 is an explanatory diagram showing an example of a procedure C in the flow chart of the insert operation to the duplicated ID sequential arrangement table according to the first embodiment of the present invention;

FIG. 17 is an explanatory diagram showing an example of a procedure D in the flow chart of the insert operation to the duplicated ID sequential arrangement table according to the first embodiment of the present invention;

FIG. 18 is an explanatory diagram showing an example of a procedure E in the flow chart of the insert operation to the duplicated ID sequential arrangement table according to the first embodiment of the present invention;

FIG. 19 is an explanatory diagram showing an example of a structure of a table including first-order and second-order duplicated keys according to a second embodiment of the present invention and a structure thereof on storage space;

FIG. 20 is an explanatory diagram showing a flow chart of insert operation to a duplicated ID sequential arrangement table according to the second embodiment of the present invention;

FIG. 21 is an explanatory diagram showing an example of a procedure A in the flow chart of the insert operation to the duplicated ID sequential arrangement table according to the second embodiment of the present invention;

FIG. 22 is an explanatory diagram showing an example of a procedure B in the flow chart of the insert operation to the duplicated ID sequential arrangement table according to the second embodiment of the present invention;

FIG. 23 is an explanatory diagram showing an example of a procedure C in the flow chart of the insert operation to the duplicated ID sequential arrangement table according to the second embodiment of the present invention;

FIG. 24 is an explanatory diagram showing an example of a procedure D in the flow chart of the insert operation to the duplicated ID sequential arrangement table according to the second embodiment of the present invention;

FIG. 25 is an explanatory diagram showing a method of structuring a duplicated ID sequential arrangement table identification table, a duplicated ID sequential arrangement table ID management table and a duplicated ID sequential arrangement table parent-child relation management table suitable for operation of the duplicated ID sequential arrangement table according to the second embodiment of the present invention;

FIG. 26 is a diagram showing an example of a Z-order storage method according to a third embodiment of the present invention;

FIG. 27 is a diagram showing an example of the Z-order storage method according to the third embodiment of the present invention;

FIG. 28 is an explanatory diagram showing an example of a structure of a table including a duplicated key based on the Z-order storage method according to the third embodiment of the present invention and a structure thereof on storage space; and

FIG. 29 is an explanatory diagram showing a method of structuring a duplicated ID sequential arrangement table identification table, a duplicated ID sequential arrangement table ID management table and a duplicated ID sequential arrangement table parent-child relation management table suitable for operation of a duplicated ID sequential arrangement table according to the third embodiment of the present invention.

BEST MODE FOR CARRYING OUT THE INVENTION
First Embodiment

A first embodiment of the present invention is described with reference to FIG. 1 to FIG. 3 and FIG. 10 to FIG. 18.

FIG. 1 shows an entire structure of a database management system according to the first embodiment of the present invention. Here, a database server (1001) and an update client (1012) and a search client (1014) are connected via a network (1016).

The database server (1001) executes a database management program (1002) and includes a storage device (1008).

The database management program (1002) includes a data update module (1003) and a data search module (1004), and further includes a duplicated ID sequential arrangement table identification table (1005), a duplicated ID sequential arrangement table ID management table (1006) and a duplicated ID sequential arrangement table parent-child relation management table (1007).

The storage device (1008) stores storage space (1009), and in the storage space (1009), a duplicated ID sequential arrangement table (1010) and an index (1011) are stored.

The data update module (1003) and the data search module (1004) refer to the duplicated ID sequential arrangement table identification table (1005), the duplicated ID sequential arrangement table ID management table (1006) and the duplicated ID sequential arrangement table parent-child relation management table (1007), and identify the duplicated ID sequential arrangement table (1010) and the index (1011).

Note that, an entire system structure is the same in a second embodiment and a third embodiment described later. The duplicated ID sequential arrangement table parent-child relation management table (1007) is a table mainly used in the second embodiment and the third embodiment. In the explanation of the first embodiment described hereinafter, the duplicated ID sequential arrangement table parent-child relation management table (1007) does not appear.

A structure of the duplicated ID sequential arrangement table identification table is shown in FIG. 2. The duplicated ID sequential arrangement table identification table (1005) is a table to which a table which is an object of arrangement and management is registered so that data having duplicated ID are arranged sequentially, and is structured of a management number column (2002) and a table name column (2003). In an example of the diagram, a table named an intersection table is registered as the object of the arrangement and the management with a management number 1.

Next, a structure of the duplicated ID sequential arrangement table ID management table is shown in FIG. 3. The duplicated ID sequential arrangement table ID management table 1006 has a table name column (3002) and a column name column (3003), and a column name to be a standard of sequential arrangement is registered in correspondence with a table name which is the object of the arrangement and the management.

FIG. 10 shows an actual example of the intersection table described above. The intersection table (11001) is a table storing intersection data divided into sections shown in FIG. 9. Here, a column (11002) named ID shows a section identifier. According to a registration content of the duplicated ID sequential arrangement table ID management table (1006) mentioned above, that is, with setting an ID column value to a key, data of the intersection table (11001) is arranged and managed. In particular, storing to the duplicated ID sequential arrangement table (1010) is performed so that data of rows having the ID column value duplicated is arranged in a sequential position.

A part of a structure of the duplicated ID sequential arrangement table (1010) on the storage space is shown in FIG. 11. The duplicated ID sequential arrangement table is a table having a characteristic that a set of rows (12002) having the same ID on the storage space (1009) is arranged (12005) in adjacent segments (12004). Further, in the same manner as the case of the clustered table, empty space (12006) is secured for preparation for future insert operation.

Note that, a plurality of segments structuring a set of rows (12002) having the same ID arranged in adjacent segment is referred to shortly as a set of rows (13001) of the same ID composed of a set of segments shown in FIG. 12 in explanation hereinafter. In addition, used space (12005) and empty space (12006) are referred to shortly as used space (13002) and empty space (13003), respectively.

In a case of performing insert operation to the duplicated ID sequential arrangement table, since it is required to maintain a state in which rows having the same ID are arranged in adjacent segments physically, insertion is performed by a method different from that for a normal table. The method of performing the insert operation to the duplicated ID sequential arrangement table is explained with reference to a flow chart shown in FIG. 13.

In an insert processing (14001) performing the insert operation, firstly, size of a row to be inserted and size of empty space of the storage space are compared, and judgment (14002) whether the insertion can be performed or not is made. Here, if the insertion cannot be performed, error judgment is made (14021) and the insert processing is ended (14010). If the insertion can be performed, a position of a set of segments including a set of rows having the same ID in the storage space is searched (14003). Different from the case of the clustered table, this search (14003) is an equal condition, and there may be a case in which the set of segments cannot be found. Therefore, judgment (14004) whether the set of segments is found or not is carried out.

If the set of segments is found, judgment (14005) whether there is empty space to insert a row in a first partial set of the set of segments or not is carried out. If there is empty space, the procedure of A (14022) is carried out. In the procedure of A (14022), firstly, a lock of the index related to a corresponding table is gotten (14006), and then, the row is inserted into the empty space of the first partial set of the set of segments (14007), the index is updated (14008), the lock of the index is released (14009) and the insert processing is ended (14010). An operation example in a case where a row of ID=13 is inserted according to the procedure of A (14022) is shown in FIG. 14. Here, since there is empty space in a set of segments (15001) including a set of rows of ID=13, the row is inserted into the empty space in a set of same segments (15002).

On the other hand, if there is no empty space in the judgment (14005), sequential empty space larger than size of an entire set of rows having the same ID as that of the row to be inserted is searched in the storage space (14011), and judgment (14012) whether it is found or not is made. In a case where it is found in the judgment (14012), a procedure of B (14023) is carried out. In the procedure of B (14023), firstly, a lock related to a corresponding table is gotten (14013), then an entire existing set of rows having the same ID as that of the row to be inserted is moved (14014) to a set of new segments to be secured in the empty space found in the search (14011), the row is inserted into empty space of a set of the same segments (14015), the index is updated (14008), the lock of the index is released (14009), and the insert processing is ended (14010). An operation example in a case where a row of ID=13 is inserted according to the procedure of B (14023) is shown in FIG. 15. Here, since there is no empty space in a set of segments including a set of rows of ID=13 (16001), an existing set of rows of ID=13 is moved (16002) to a set of new segments, and the row is inserted (16003) into empty space of a set of the same segments.

In a case where sequential empty space larger than the size of the entire set of rows having the same ID as that of the row to be inserted is not found in the judgment (14012), judgment (14016) whether there is a partial set having empty space in partial sets of the set of segments having the same ID as that of the row to be inserted or not is made. In the judgment (14016), if a partial set of the set of segments having empty space is found, the procedure of C (14024) is carried out. In the procedure of C (14024), firstly, a lock related to a corresponding table is gotten (14013), the row is inserted into the partial set of the set of segments having the empty space (14018), the index is updated (14008), the lock of the index is released (14009) and the insert processing is ended (14010). An operation example in a case where a row of ID=13 is inserted according to the procedure of C (14024) is shown in FIG. 16. Here, since a set of segments including a set of rows of ID=13 is composed of two partial sets (17001, 17002), and further, empty space (17003) is smaller than size of an entire set of rows of ID=13, the row is inserted (17004) into a second partial set (17002).

On the other hand, in a case where a partial set of segments having empty space is not found in the judgment (14016), a procedure of D (14025) is carried out. In the procedure of D (14025), firstly, a lock related to a corresponding table is gotten (14019), and then, the row is inserted into a new partial set of the set of segments (14020), the index is updated (14008), the lock of the index is released (14009) and the insert processing is ended (14010). An operation example in a case where a row of ID=13 is inserted according to the procedure of D (14025) is shown in FIG. 17. Here, since there is no empty space in a set of segments (18001) including a set of rows of ID=13, and further, empty space (18002) is not large enough, a new partial set (18003) is secured in the empty space (18002) and the row of ID=13 is inserted (18004).

Further, in a case where a set of segments including a set of rows having the same ID as that of the row to be inserted is not found in the judgment (14004), a procedure of E (14026) is carried out. In the procedure of E (14026), firstly, a lock related to a corresponding table is gotten (14019), and then, the row is inserted (14020) into a new partial set of the set of segments, the index is updated (14008), the lock of the index is released (14009) and the insert processing is ended (14010). An operation example in a case where a row of ID=13 is inserted according to the procedure of E (14025) is shown in FIG. 18. Here, since there is not a set of segments including a set of rows of ID=13, a new partial set (19002) is secured in the empty space (19001), and the row of ID=13 is inserted (19003).

In the present embodiment, by performing the procedure of C (14024), the rows having duplicated ID are stored in non-sequential space. However, by performing the procedure of B (14023) in later insert operation, the rows are rearranged to sequential space.

Second Embodiment

By arranging a plurality of columns among columns structuring a table at physical vicinity on storage, I/O to a storage device may be sped-up. FIG. 19 shows an example in which among columns structuring an intersection table (21001), an ID column (21003) and a road type column (21002) are arranged (21005, 21006) at physical vicinity on storage space (21004).

In the present embodiment, a plurality of columns designated to be arranged at physical vicinity are sequentially called n-th-order key column, and in the more dominant column, a value of n is made larger. In this example, in a case of an equal condition search of the ID column (21003) which is a second-order key column, all data can be fetched collectively. For example, as for an equal condition search having “ID is equal to 2” as a search condition, a set of segments (21013) of ID=2 can be fetched collectively, and as for an equal condition search having “ID is equal to 3” as a search condition, a set of segments (21014) with ID=3 can be fetched collectively.

Further, also in a case of an equal condition search of a road type column (21002) which is a first-order key column, since rows matching the condition are rather organized, data can be fetched at high speed. For example, as for an equal condition search having “road type is equal to highway” as a search condition, a plurality of sets of segments (21007, 21010) of road type=highway are fetched, in a case of “road type is equal to national road”, a plurality of sets of segments (21008, 21011) of road type=national road are fetched, and in a case of “road type is equal to prefectural road”, a plurality of sets of segments (21009, 21012) of road type=prefectural road are fetched. Such a table is one obtained by expanding a duplicated ID sequential arrangement table so as to cope with a plurality of columns.

In a case where insert operation is performed to the duplicated ID sequential arrangement table expanded to a plurality of columns, it is required to maintain a state in which rows having the same ID are arranged in adjacent segments physically over a plurality of columns, and therefore, insertion is performed by a method different from that for a normal table.

A method of performing insert operation to the duplicated ID sequential arrangement table expanded to a plurality of columns is described with reference to a flow chart shown in FIG. 20.

In an insert processing (22001) performing the insert operation, firstly, size of a row to be inserted and size of empty space of storage space are compared, and judgment (22002) whether the insertion can be performed or not is made. Here, if the insertion cannot be performed, error judgment is made (22027) and the insert processing is ended (22014). If the insertion can be performed, an insert portion is searched in the storage space. For example, in an example shown in FIG. 19, the insert portion of a row of ID=2 and road type=national road is one (21008) included in a set of segments of ID=2 among a plurality of sets of segments (21008, 21011) of road type=national road. Here, using a counter variable i, a position of an i-th-order set of segments to be the insert portion is searched (22004) sequentially from i=1 (22003). And then, judgment (22005) whether it is found or not is made, and if it is not found, 1 is added to i (22006) and the search (22004) is repeated. Here, in a case where i reaches a highest order, the procedure goes ahead in the judgment (22005).

Next, judgment (22007) whether there is empty space in a first partial set of an i-th-order set of segments found in the judgment (22005) or i is a highest order or not is made, and if one of these conditions is satisfied, a lock of an index related to a corresponding table is gotten (22008), and then, judgment (22009) whether i=1 is true or not is made. If i=1 is true in the judgment (22009), a procedure of A (22028) is carried out. In the procedure of A (22028), firstly, the row is inserted into an empty space of a first-order set of segments (22011), the index is updated (22012), the lock of the index is released (22013) and the insert operation is ended (22014). An operation example in a case where a row of ID=2 and road type=national road is inserted according to the procedure of A (22028) is shown in FIG. 21. Here, since there is empty space in a set of segments (23001) including a set of rows of ID=2 and road type=national road, the row is inserted (23002) into the empty space of a set of the same segments.

On the other hand, in a case where i=1 is not true in the judgment (22009), a procedure of B (22029) is carried out. In the procedure of B (22029), since it is a case in which a first-order set of segments to which the row is inserted is not found in the judgment (22004), firstly, empty space is secured (22010) for the first-order set of segments. And then, the row is inserted (22011) into the empty space of the first-order set of segments, the index is updated (22012), the lock of the index is released (22013), and the insert operation is ended (22014). An operation example in a case where a row of ID=2 and road type=national road is inserted according to the procedure of B (22029) is shown in FIG. 22. Here, although there is no set of segments including a set of rows of ID=2 and road type=national road, there is sequential empty space (24002) just after a set of segments (24001) including a set of rows of ID=2, and therefore, space for the first-order set of segments is secured just after the set of segments (24001) including the set of rows of ID=2, and the row is inserted (24003) thereinto.

In a case where both the conditions are not satisfied in the judgment (22007), an action moving a second or higher order set of segments to other empty space is carried out. Here, using a counter variable j, sequential empty space larger than a j-th-order set of segment is searched (22016) sequentially from j=highest order (22015). And then, judgment (22017) whether it is found or not is made, and if it is not found, 1 is subtracted from j (22018), and the search (22016) is repeated.

Here, in a case where j reaches 0, the procedure goes ahead in the judgment (22017). And then, the lock of the index related to the corresponding table is gotten (22019).

Next, judgment (22020) whether j=0 is true or not is made. In a case where j=0 is true in the judgment (22020), a procedure of C (22030) is carried out. In the procedure of C (22030), since j=0 is true in the judgment (22017), that is, sequential empty space larger than the first-order set of segments cannot be secured, the first-order set of segments is divided. Firstly, empty space for the first-order set of segments is secured (22021), the row is inserted (22022) into the empty space of the first-order set of segments, the index is updated (22012), the lock of the index is released (22013) and the insert operation is ended (22014). An operation example in a case where a row of ID=2 and road type=national road is inserted according to the procedure of C (22030) is shown in FIG. 23. Here, since there is no empty space in the first-order set of segments (25001) including a set of rows of ID=2 and road type=national road and size of empty space (25002) is smaller than the first-order set of segments (25001), empty space for the first-order set of segments is secured in the empty space (25002), and then, the row is inserted (25003) into the empty space of the first-order set of segments.

In a case where j=0 is not true in the judgment (22020), a procedure of D (22031) is carried out. In the procedure of D (22031), since judgment that sequential empty space larger than a j-th-order set of segments higher than first-order is found in the judgment (22017) is made, the j-th-order set of segments is rearranged in a sequential space.

Firstly, sequential empty space larger than the j-th-order set of segments is secured (22023) in the empty space found in the search (22016), and then, the first-order set of segments is moved (22024) to n which is an insertion row in ascending order, the row is inserted (22025) into empty space of the first-order set of segments, the first-order set of segments is moved (22026) from n which is the insertion row or later in ascending order, the index is updated (22012), the lock of the index is released (22013), and the insert operation is ended (22014). An operation example in a case where a row of ID=2 and road type=national road is inserted according to the procedure of D (22031) is shown in FIG. 24. Here, since there is no empty space in a first-order set of segments (26001) including a set of rows of ID=2 and road type=national road and size of empty space (26003) is larger than a second-order set of segments (26002), the second-order set of segments is rearranged (26004) collectively in sequential space.

Finally, a duplicated ID sequential arrangement table identification table, a duplicated ID sequential arrangement table ID management table and a duplicated ID sequential arrangement table parent-child relation management table according to the present embodiment are shown in FIG. 25. The duplicated ID sequential arrangement table identification table (27001) is composed of a management number (27002) column and a table name (27003) column, and is the same as the duplicated ID sequential arrangement table identification table 1005 according to the first embodiment shown in FIG. 2. The duplicated ID sequential arrangement table ID management table (27004) is composed of a table name (27005) column and a column name (2700.6) column and the duplicated ID sequential arrangement table parent-child relation management table is composed of a table name (27008) column, a column name (27009) column and an order number (27010) column.

Here, since the intersection table (21001) is stored as the duplicated ID sequential arrangement table with designating the road type column (21002) in FIG. 19 as a first-order key column and the ID column (21003) as a second-order key column, a row (27011) indicating the intersection table is included in the duplicated ID sequential arrangement table identification table (27001), a row (27012) showing the road type column of the intersection table and a row (27013) indicating the ID column are included in the duplicated ID sequential arrangement table ID management table (27004), and further, a row (27014) showing that an order number of the road type column is 1 and a row (27015) showing that an order number of the ID column is 2 are included in the duplicated ID sequential arrangement table parent-child relation management table (27007).

Third Embodiment

As an effective storage method of map data divided into sections (10001) shown in FIG. 9, there is a Z-order storage method. An example of the Z-order storage method is shown in FIG. 26. The Z-order storage method is a method allocating identification numbers of sections in Z-shaped order, and storing in order of the section numbers.

As an example of an allocation method of the identification numbers, in FIG. 26, identification numbers from 1 to 4 are allocated in Z-shaped (28005) order to a set of adjacent four sections (28001, 28002, 28003, 28004) at left top. By using the Z-order storage method, in a case where the intersection list fetch is performed for adjacent nine sections around own car marked as A (28006), intersection data of adjacent four sections marked as a (28007) can be fetched collectively.

The Z-order storage method can be expanded to space of adjacent four or more sections. As an example thereof, an example of second-order Z-order storage handling adjacent 16 sections is shown in FIG. 27. A set of the adjacent 16 sections is structured by allocating four sets of adjacent four sections in Z-shaped order. In FIG. 27, identification numbers are allocated to four sets of adjacent four sections (29001, 29002, 29003, 29004) in Z-shape (29005). Here, sets of adjacent four sections (29001, 29002, 29003, 29004) are referred to as a first-order Z-shape, and a set of adjacent 16 sections (29005) is referred to as a second-order Z-shape.

By allocating identification numbers in second-order Z-shaped order, sections from 1 (29006) to 16 (29007) can be stored so as to be fetched collectively. And, by enlarging the Z-shape in the same manner, expansion obtaining such as a third-order Z-shape composed of adjacent 64 sections or a fourth-order Z-shape composed of adjacent 256 sections can be made.

The four sections included in the same first-order z-shape have a characteristic that values obtained by dividing (section number−1) by four are equal. The value obtained by adding 1 to this quotient is defined as a first-order Z-shape identifier. The first-order Z-shape identifiers (28008, 28009, 28010, 28011, 28012, 28013, 28014, 28015, 28016, 28017, 28018, 28019, 29020, 28021, 28022, 28023) are identifiers allocated according to the above definition.

In the same manner, as for the second-order Z-shape, one obtained by adding 1 to a quotient obtained by dividing (first-order Z-shape identifier—1) by four is referred to as a second-order Z-shape identifier. The numeric FIGS. 1, 2, 3, 4 indicated by reference symbols 29008 to 29011 in FIG. 27 are the second-order Z-shape identifiers allocated sequentially according to the above definition.

The Z-order storage method handled in the present embodiment can be embodied as a special example of the duplicated ID sequential arrangement table expanded to a plurality of columns shown in the second embodiment irrespective of a storage order.

FIG. 28 shows an example in which a first-order Z-order storage method is expanded to a plurality of columns. Here, an example in which among columns structuring the intersection table (30001), four sections adjacent in first-order Z-shaped order with regard to an ID column (30002) are arranged (30014, 30015) at physical vicinity on storage space (30003) is shown. Here, among the four sections adjacent in first-order Z-shaped order shown in FIG. 26, a fourth adjacent four sections (28011) and a fifth adjacent four sections (28012) are arranged. The fourth adjacent four sections (28011) are a set of segments (30014) of first-order Z-shaped order identifier=4 and four sections (30006, 30007, 30008, 30009) structuring this are arranged in sequential space as a first-order key column (30004) on the storage space.

And, the fifth adjacent four sections (28012) are a set of segments (30015) of first-order Z-shaped order identifier=5 and four sections (30010, 30011, 30012, 30013) structuring this are arranged in sequential space as a first-order key column (30004) on the storage space (30003) in the same manner. As described above, by regarding the first-order Z-shape identifiers as the second-order key columns, it can be embodied as a special example of the duplicated ID sequential arrangement table expanded to two columns. Recursively, by regarding a n-th Z-shape identifiers as a n+1-th key columns, it can be embodied as a special example of a duplicated ID sequential arrangement table expanded to n+1 columns.

A duplicated ID sequential arrangement identification table, a duplicated ID sequential arrangement ID management table and a duplicated ID sequential arrangement table parent-child relation management table used in the present embodiment are described with reference to FIG. 29. In the present embodiment, in comparison with the case in the second embodiment, information to be managed in the duplicated ID sequential arrangement table parent-child relation management table is different.

In the duplicated ID sequential arrangement table parent-child relation management table (31007) used in the present embodiment, in addition to a table name (31008) column, a column name (31009) column and an order number (31011) column, a condition column (31010) column is provided so that it can be identified whether it follows to the Z-order storage method or not. The duplicated ID sequential arrangement table identification table (31001) is composed of a management number (31002) column and a table name (31003) column in the same manner as in the case of the second embodiment, and the duplicated ID sequential arrangement table ID management table (31004) is structured of a table name (31005) column and a column name (31006) column also in the same manner as in the case of the second embodiment.

Here, in order to arrange four vicinities adjacent in the first-order Z-shaped order adjacently with respect to the ID column (30002) of the intersection table (30001) in FIG. 28, a row (31012) indicating the intersection table is included in the duplicated ID sequential arrangement table identification table (31001), a row (31013) indicating the ID column of the intersection table is included in the duplicated ID sequential arrangement table ID management table (31004) and a row (31014) showing that an order number of the ID column is 1 and the Z-order storage method is applied and a row (31015) showing that an order number of the ID column is 2 and the Z-order storage method is applied in the same manner are included in the duplicated ID sequential arrangement table parent-child relation management table (31007).

Note that, in the present embodiment, the Z-order storage method is taken as an example, however, an N-order storage method arranging N-shaped order four vicinities in the same manner and a clock-order storage method arranging four vicinities clockwise can known easily by analogy. Also in the duplicated ID sequential arrangement table parent-child relation management table (31007) in the present embodiment, by designating the N-order storage method or the clock-order storage method in the condition column (31010), the N-order storage method or the clock-order storage method can be realized.

INDUSTRIAL APPLICABILITY

In embedded applications, since restriction of usable memory size is severe and a large buffer cannot be allocated like in DBMS for server, I/O performance of a storage device affects directly upon total search performance. Especially in car navigation system application, since a file structure optimized so that a conventional car navigation system can use sequential access of an optical disk device efficiently is used, even if a hard disk device of higher-speed is introduced, by occurrence of a random access, performance is degraded with respect to an optical disk. Therefore, it is necessary to perform a control so as to restrict the random access using the present invention.

Database Management Program

Information

Publication Number

Date Filed

Date Published

Inventors

CPC

US Classifications

International Classifications

Abstract

Description

Claims

PCT Information