JP2011048801A - Bit stream data sorting device, method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an efficient sorting method which does not cause invalid processing in sort processing of bit string data. <P>SOLUTION: A key that is a first reference is selected from keys to be sorted, and for each of a group of keys smaller than the selected key and a group of keys larger than the selected key, bit string comparison of the reference key with each of the keys to be sorted, the keys being composed of a bit string, is performed to determine a differential bit position that is a bit position where a different bit value is obtained first. The keys to be sorted are classified by the differential bit position, and the classification by the differential bit position with respect to a referential key among a plurality of keys classified into the same differential bit position is further repeated, whereby a sorted key sequence is obtained. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a bit string data sorting apparatus, method, and program.

  In recent years, with the progress of informatization of society, large-scale databases are being used in various places. In order to search for a record from such a large database, it is usual to search for an item in the record associated with the stored address of each record using an index key to find a desired record. A character string in full-text search can also be regarded as a document index key.

Since these index keys are expressed by bit strings, it can be said that a database search is reduced to a bit string search.
On the other hand, as a process related to the database, a sort process based on an index key of records in the database is performed. This sort process is also reduced to a bit string sort process. In the following description, the bit string data may be simply referred to as a key.

  Various sort methods have been developed, and the following Patent Document 1 introduces quick sort, radic sort (radix sort), and the like. Patent Document 2 also describes radix sort.

  FIG. 1 is a diagram for explaining the concept of a conventional radix sort. According to the radix sort, the key which is the 4-bit bit string to be sorted illustrated in FIG. 1 is sorted by repeating the classification by the bit value at each bit position from the 0th bit to the 3rd bit. The Hereinafter, the concept of the radix sort will be described with reference to FIG.

FIG. 1 shows a key string 100 including keys to be sorted. In the example of FIG. 1, the key “1010” is stored in the storage area in which the key position 101 that is the position of the key included in the key string 100 is 100a (hereinafter referred to as the key position 100a). Exists. Further, the keys “0111”, “1000”, “0010”, “0110” are respectively provided in the key positions 100b, 100c, 100d, 100e, 100f, and 100g.
, “1111”, “1011” exist.

As shown in FIG. 1, first, classification by the 0th bit of the key included in the key string 100 is performed by the classification (0th bit) 110a for each bit position. As a result, a set 120a of the 0th bit value 0 consisting of the key "0111", the key "0010", and the key "0110", the key "1010", the key "1000", and the key "1111"
And a set 121a of value 1 consisting of the key “1011” is obtained. Next, the set 0a of the value 0 is classified by the value of the first bit by the classification (first bit) 110b for each bit position, and the set 121a of the value 1 is classified by the value of the first bit by the classification (first bit) 110e for each bit position. Done.

  In the classification (first bit) 110b for each bit position, one key “0010” is obtained as the set 120b of the first bit value 0, so the key position 131 that is a position for storing the key of the sorted key string 130. Is stored in the storage area 130a (hereinafter referred to as a key position 130a). Further, a set of the key “0111” and the key “0110” is obtained as the value 1 set 121b of the first bit, and the classification (second bit) 110c for each bit position is based on the value of the second bit. Classification is performed.

In the second bit classification 110c, since the second bits of the key “0111” and the key “0110” are both 1, only the first bit value set 121c is obtained, and the first bit value 1 The same key as the set 121b is the target of classification by the third bit, and classification by the third bit is performed by the classification (third bit) 110d for each bit position. In the classification (third bit) 110d for each bit position, one key “0110” is obtained as the set 120d of the third bit value 0, and is stored in the key position 130b. Similarly, since one key “0111” is obtained as the set 121d of the value 1 of the third bit, it is stored in the key position 130c for storing the key of the next value in the sorted key string 130. Note that the sorted key string is stored from the smaller key in the code order of the key position.

On the other hand, in the classification for each bit position (first bit) 110e, the key “1010” and the key “1000” are set as the set 120e of the first bit value 0.
, A set of keys “1011” is obtained, and classification based on the value of the second bit is performed in the classification of each bit position (second bit) 110f. In addition, since one key “1111” is obtained as a set 121e of the value 1 of the first bit, it is stored in the key position 130g of the sorted key string 130.

  In the classification (second bit) 110f for each bit position, one key “1000” is obtained as a set 120f of the second bit value 0, and is stored in the key position 130d of the sorted key string 130. Further, a set of the key “1010” and the key “1011” is obtained as the set 121f of the value 1 of the second bit, and is based on the value of the third bit in the classification of each bit position (third bit) 110g. Classification is performed.

In the classification (third bit) 110g for each bit position, one key “1010” is obtained as a set 120g of the value 0 of the third bit, and is stored in the key position 130e of the sorted key string 130. Similarly, since one key “1011” is obtained as the set 121g of the value 1 of the third bit, it is stored in the key position 130f that is the next storage position of the sorted key string 130.

  Through the above processing, the keys of the key string 100 are sorted and stored in the key positions 130a to 130g of the sorted key string 130. However, when sorting is performed by the above-described radix sort method, invalid processing in which classification is not performed occurs as seen in the classification (second bit) 110c for each bit position shown in FIG.

JP 2002-116907 A JP 2005-316663 A

Therefore, the problem to be solved by the present invention is to provide an efficient sorting method in which invalid processing does not occur in the sort processing of bit string data.

  According to the sorting process of the present invention, an arbitrary reference key is selected as a reference key from the sort target key that is a key consisting of a bit string to be sorted, the size of the reference key and the sort target key is compared, and the sort target key For each group of keys larger than the reference key and keys smaller than the reference key, the bit string comparison of the reference key and the sort target key is performed, and a difference bit position that is a bit position that first becomes a different bit value is obtained. The sort target keys are classified according to the position, and the sorted key string is obtained by further repeating the classification based on the difference bit position for the reference key among the plurality of keys classified into the same difference bit position.

  According to the present invention, since the keys are sorted by the difference bit position, invalid processing that does not perform classification does not occur. Therefore, efficient sort processing can be realized.

It is a figure explaining the concept of the conventional radix sort. It is a figure explaining the concept of the sort process by the difference bit position in one embodiment of this invention. It is a figure explaining the example of a data structure used by the sort process by the difference bit position in one embodiment of this invention. It is a figure explaining the functional block structural example of the bit string data sorting apparatus in one embodiment of this invention. It is a figure explaining the hardware structural example for implementing this invention. It is a figure explaining the example of a processing flow of the whole sort process in one embodiment of this invention. It is a figure explaining the example of a processing flow which classify | categorizes according to a difference bit position about the group of the sort key smaller than the reference | standard key and the group of a big sort key in one embodiment of this invention. It is a figure explaining the example of a process flow of the front | former stage of the process which classifies the group of the sort key larger than the reference | standard key in one embodiment of this invention. It is a figure explaining the example of a processing flow of the latter part of processing which classifies a group of sort keys larger than a standard key in one embodiment of the present invention. It is a figure explaining the example of a process flow of the front | former stage of the process which classifies | sorts the group of the sort key smaller than the reference | standard key in one embodiment of this invention. It is a figure explaining the example of a processing flow of the latter part of processing which classifies a group of sort keys smaller than a standard key in one embodiment of the present invention. It is a figure explaining the example of a processing flow of the process which takes out the sort key smaller than the reference | standard key in order of difference bit position in one embodiment of this invention. It is a figure explaining the example of a processing flow of the process which subdivides the some sort key linked with the difference bit position in one embodiment of this invention in order of a difference bit position. It is a figure explaining the example of a processing flow of the process which takes out the sort key larger than the reference | standard key in order of difference bit position in one embodiment of this invention. It is a figure explaining the example of a processing flow of the process which subdivides the some sort key linked with the difference bit position in one embodiment of this invention in order of a difference bit position. It is a figure explaining the flow of the data which takes out the key in descending order in one embodiment of this invention.

Hereinafter, a mode for carrying out the present invention will be described by exemplifying the same key as the key included in the key column 100 shown in FIG. 1 as a sort target key.
In Japanese Patent Application No. 2008-325692, the present applicant has already proposed a classification method based on the difference bit position of the sort target key. This proposed classification method uses the reference key as the minimum value or the maximum value of the sort target keys.
On the other hand, in the embodiment of the present invention described below, an arbitrary one can be selected from the sort target keys as the reference key. For example, the sort target key stored at the top position of the storage area in which the sort target key is stored is used as the reference key.

FIG. 2A is a diagram for explaining the concept of the sorting process based on the difference bit position according to the embodiment of the present invention. The sort target key string 100 and the sorted key string 130 are the same as those shown in FIG. The key “1010” located at the top key 100a of the key string 100 is selected as the reference key 148a.
Each key constituting the key sequence 100 is different from the reference key 148a and the reference key in the classification 149a based on the difference bit position, and the difference bit position from the reference key 148a and the key group 140a and the difference bit position from the reference key. 2 is classified into a key group 142a which is larger than the reference key, a key group 141a whose difference bit position is 1 from the reference key 148a, and a key group 143a whose difference bit position is 3 from the reference key.

In the example of FIG. 2A, the reference key 148a is “1010”. The key group 140a includes keys “0111”, “0010”, and “0110”, and the key group 142a includes only the key “1000”. The key group 141a includes only the key “1111”, and the key group 143a includes only the key “1011”.

The reference key 148a is stored in the key position 130e of the sorted key row 130 as indicated by the dotted arrow 158a in the figure. Further, since there is only one key in the key group 142a, the key group 142a is stored in the key position 130d of the sorted key string 130 as indicated by the dotted arrow 152a in the figure. Similarly, since there is only one key in the key group 141a and one key in the key group 143a, the keys are stored in the key positions 130g and 130f of the sorted key row 130, as indicated by the dotted arrows 151a and 153a in the figure. Is done.
The storage of keys in the sorted key string 130 will be described in detail later.

Since the key group 140a whose difference bit position with respect to the reference key 148a is 0 includes a plurality of keys, the key group 140a in the classification 149b based on the difference bit position as shown by the dotted arrow set 150a in the figure. The key group 141b has a difference bit position of 1 between the maximum value key 148b and the maximum value key 148b, and the key group 143b has a difference bit position of 3 from the maximum value key 148b. In the illustrated example, the maximum value key 148b is “0111”. The key group 141b includes only the key “0010”, and the key group 143b includes only the key “0110”.

The maximum value key 148b is stored in the key position 130c of the sorted key row 130 as indicated by the dotted arrow 158b in the figure. Since the key group 141b whose difference bit position with respect to the maximum value key 148b is 1 is one key, it is stored in the key position 130a of the sorted key string 130 as indicated by the dotted arrow 151b in the figure. Further, since there is one key included in the key group 143b whose difference bit position is 3 with respect to the maximum value key 148b, the fine classification is finished, and as shown by the dotted arrow 153b in the figure, the keys of the sorted key sequence 130 Stored at position 130b.

With the above processing, the keys of the key string 100 are stored in ascending order at the key positions 130a to 130g of the sorted key string 130, and the sorting process is completed. Since the sorting process according to the present embodiment repeats the classification based on the difference bit position, the key classification is always executed for each classification process.

Note that the key positions 130a to 130g of the sorted key string 130 do not have to be continuous storage area addresses. The key positions 130a to 130d, 130e, and 130f to 130g may be addresses of three separate storage areas. Further, for example, the key positions 130a to 130d may be assigned in the direction from the largest address of the storage area to the smallest address. If the key positions 130e to 130d of the key that is smaller than the reference key and the direction of the address and the key positions 130f to 130g of the key that is larger than the reference key and the direction of the address are known, the keys are arranged in ascending order. Or it is easy to take out in descending order. Of course, it is also possible to assign the key positions 130a to 130g in the order of the addresses of the continuous storage areas.

  Next, the data structure and functional block configuration for realizing the sort processing based on the difference bit position in the present embodiment whose concept has been described with reference to FIG. 2A will be described. The sort processing based on the difference bit positions in the present embodiment can be realized by preparing the function blocks for executing the processing of the classifications 149a and 149b based on the difference bit positions illustrated in FIG. 2A for all combinations of keys to be sorted. Obviously. However, this is a waste of resources, and the realization method is not realistic. Therefore, in the present embodiment, the following devices are devised.

  FIG. 2B explains an example of a data structure used in the sort processing based on the difference bit position according to the embodiment of the present invention. As shown in FIG. 2B, a key table 309, a difference bit position table 310a, a difference bit position table 310b, a link table 311 and a reference key 650a which is a storage area for setting a reference key are used.

  The key table 309 is set by reading the sort target key 602 when there is a sort request. In the example of FIG. 2B, as the seven keys 602 to be sorted, those included in the key string 100 shown in FIGS. 1 and 2A are read. The head of the reading position 601 in the key table 309 is P0, and is followed by P1, P2, P3, P4, P5, and P6. The key set at the reading position P0 is used as the reference key, but the key set at another reading position can be used as the reference key. As indicated by the dotted arrow 689d, the reference key is set to the reference key 650a.

  In the example of FIG. 2B, the difference bit position table 310a is smaller than the reference key and has the same difference bit position with respect to the reference key. In the example of FIG. 2B, the maximum bit read position 601 is the parent link 612a. Store every time. In the example of FIG. 2B, as indicated by a dotted arrow 660a, the largest key “0111” of the keys “0111”, “0010”, and “0110” whose difference bit position from the reference key “1010” is 0 is displayed. The read position P1 is stored at a position where the difference bit position 611a of the difference bit position table 310a is 0. Also, as indicated by the dotted arrow 662a, the key whose difference bit position is 2 with respect to the reference key “1010” is only the key “1000”, so the read position P2 of the key “1000” is the difference bit position table 310a. Are stored in positions where the difference bit position 611a is 2.

The difference bit position table 310b stores the smallest key reading position 601 among the keys that are larger than the reference key and have the same difference bit position with respect to the reference key as the parent link 612b for each difference bit position 611b. is there. In the example of FIG. 2A, as indicated by the dotted arrow 661a, the key whose difference bit position is 1 with respect to the reference key “1010” is only the key “1111”, so that the reading position P5 of the key “1111” is the difference. The difference bit position 611b of the bit position table 310b is stored at a position of 1. Further, as indicated by the dotted arrow 663a, the key whose difference bit position is 3 from the reference key “1010” is only the key “1011”, and therefore the key “1011”.
Read position P6 is stored at a position where the difference bit position 611b of the difference bit position table 310b is 3.

  The link table 311 is for accessing a key having the same differential bit position with respect to the reference key. As shown in the figure, a link 622 indicating a key reading position having the same differential bit position is stored with respect to the key reading position 621.

As indicated by the dotted arrow 671b from the difference bit position table 310a to the link table 311, the read position P1 is set to the read position P1 (parent link) of the maximum value key of the key having the same difference bit position 0. The maximum value key “0111” of the key having the difference bit position 0 with respect to the reference key “1010” at the reading position of the link table 311 as the reading position 621 (hereinafter referred to as the reading position P1 of the link table 311). The read position P4 of the key table 309 of the key “0110” having the same differential bit position 0 as “” is stored.

As indicated by a dotted arrow 674c in the figure, the reading position P3 of the key table 309 of the key “0010” having the same differential bit position 0 is stored at the reading position P4 of the link table 311. Further, as indicated by a dotted arrow 673c in the figure, the reading position P3 of the key table 309 having the same value is stored in the reading position P3 of the link table 311. This indicates that there is no more key having the same differential bit position 0.

In addition, as indicated by a dotted arrow 672b from the difference bit position table 310a to the link table 311, the link table 311 with respect to the reading position P2 (parent link) of the maximum value key of the key having the same difference bit position 2 The read position P2 of the key table 309 having the same value as the read position P2 of the link table 311 is stored in the read position P2. This indicates that the key “1000” at the reading position P2 in the key table 309 is the largest key of the key whose difference bit position is 2 from the reference key “1010” and is unique.

On the other hand, as indicated by a dotted arrow 675b from the difference bit position table 310b to the link table 311, with respect to the reading position P5 (parent link) of the minimum value key of the key having the same difference bit position 1 with the reference key. The read position P5 of the key table 309 having the same value as the read position P5 of the link table 311 is stored in the read position P5 of the link table 311. This indicates that the key “1111” at the reading position P5 in the key table 309 is the smallest key of the key whose difference bit position is 1 from the reference key “1010” and is unique.

Similarly, as indicated by a dotted arrow 676b from the difference bit position table 310b to the link table 311, the link table for the minimum value key reading position P5 of the key having the same difference bit position 3 with the reference key is shown. The reading position P5 of 311 stores the reading position P6 of the key table 309 having the same value as the reading position P6 of the link table 311. This indicates that the key “1011” at the reading position P6 in the key table 309 is the smallest key of the key whose difference bit position is 3 from the reference key “1010” and is unique.

The states of the difference bit position tables 310a and 310b and the link table 311 shown in FIG. 2B described above are obtained by executing the classification 149a based on the difference bit positions shown in FIG. 2A.
As is apparent from the above description, the key table 309 described above is an example of the sort target key storage means of the invention according to claim 1. The difference bit position tables 310a and 310b and the link table 311 constitute one embodiment of the classification target key difference bit position storage means, and the read position 601 of the key table 309 corresponds to key identification information.

FIG. 2C is a diagram illustrating a functional block configuration example of the bit string data sorting device according to one embodiment of the present invention.
As shown in the figure, the bit string data sorting apparatus 200 includes a grouping unit 205, a difference bit position calculation unit 210, a difference bit position classification unit 220, a sorted key output unit 230, and a control unit 240. Hereinafter, an outline of the operation of each unit will be described with reference to the example of FIG. 2B.

The grouping unit 205 compares the size of all sort target keys stored in the key table 309 with the reference key, and performs grouping of a group of keys smaller than the reference key and a group of keys larger than the reference key. The difference bit position calculation means 210 calculates the difference bit position of the key to be classified by the difference bit position and the reference key for calculating the difference bit position among the keys 602 stored in the key table 309. . The first reference key for calculating the difference bit position is the reference key.
The difference bit position classification means 220 writes the difference bit position table 310a and the link table 311 based on the first calculation result of the difference bit position calculation means 210 for the key group smaller than the reference key, and is larger than the reference key. Based on the first calculation result of the difference bit position calculation unit 210 for the key group, the difference bit position table 310b and the link table 311 are written to perform classification based on the difference bit position of the key.
As a result of the first classification, a group of keys smaller than the reference key has a key identification information as a reading position 601 in the key table 309, a set of difference bit positions of 0 (P1, P4, P3) and a difference bit position of 2 (P2). Key groups larger than the reference key are classified into a pair (P5) with a difference bit position of 1 and a set (P6) with a difference bit position of 3.

  The control unit 240 groups the keys stored in the key table 309 into a group of keys smaller than the reference key and a larger group by the grouping unit 205. The target of the first classification process is set to all keys smaller than the reference key stored in the key table 309 and all keys larger than the reference key. In this case, the reference key serving as a reference for calculating the difference bit position is the key “1010” stored at the reading position P0. The sorted key output unit 230 outputs the reference key “1010” to the sorted key table 313.

  The control unit 240 performs control so that the difference bit position calculation by the difference bit position calculation unit 210 and the classification process by the difference bit position classification unit 220 are repeated for a set including a plurality of keys after classification.

FIG. 3 is a diagram for explaining a hardware configuration example for carrying out the present invention.
Sort processing according to the present invention is performed by a data processing device 301 including at least a central processing unit 302 and a cache memory 303 using a data storage device 308. The data storage device 308 having the key table 309, the difference bit position table 310a, the difference bit position table 310b, the link table 311 and the sorted key table 313 can be realized by the main storage device 305 or the external storage device 306, or It is also possible to use a remote device connected via the communication device 307.

  In the example of FIG. 3, the main storage device 305, the external storage device 306, and the communication device 307 are connected to the data processing device 301 by a single bus 304, but the connection method is not limited to this. In addition, the main storage device 305 can be in the data processing device 301, the key table 309 is in the external storage device 306, and other tables are in the main storage device 305. It is obvious that the hardware configuration can be appropriately selected according to the size of the set of sorting target keys.

  Although not particularly illustrated, it is natural that the temporary storage area of the main storage device 305 corresponding to each process is used in order to use various values obtained during the process in the subsequent process. . In the following description, a value stored or set in the primary storage area may be referred to as a temporary storage area name.

  Next, an outline of the sorting process according to the embodiment of the present invention will be described with reference to FIG. Moreover, in the description, it demonstrates using the illustration of FIG. 2A or FIG. 2B suitably.

  FIG. 4 is a diagram for explaining an example of the processing flow of the entire sort processing in one embodiment of the present invention. It is assumed that the key to be sorted is stored in the key table. That is, the keys included in the key string 100 shown in FIG. 2A are stored in the key table 309 as shown in FIG. 2B.

First, in step S401, the head position of the key table is set to a key table reading position (hereinafter, simply referred to as a reading position). The “reading position of the key table” here is one of the “temporary storage areas corresponding to each processing in order to use various values obtained during the processing in the subsequent processing” described above. . In the following description, it is sometimes omitted that “the key table read position” is a primary storage area (not shown), such as “the head position of the key table is set as the key table read position”. .
With the above setting, in the example of FIG. 2B, P0 is set at the reading position of the key table.

In step S402, the key indicated by the head position is extracted from the key table as a reference key. Here, “take out as a reference key” means to set a reference key which is a temporary storage area. This temporary storage area is not shown in most cases. In the following, similar notation may be used.

  Proceeding to step S403, the next reading position is set as the reading position. Next, in step S404, it is determined whether all the keys have been processed, that is, whether the classification process at the first stage for all the keys has been completed.

If all the keys have not been processed, the process proceeds to step S405, where the key pointed to by the reading position is extracted from the key table as a sort key, and the reading position is set as the reading position of the sort key. In step S406, a difference bit position between the sort key and the reference key is obtained, the sort key is classified based on the difference bit position, and the process returns to step S403.
In the process of step S406, the difference bit position table and the link table are set as the classification process in the first stage. Details of the processing will be described later with reference to FIGS. 5, 6A and 6B, and FIGS. 7A and 7B.

On the other hand, if it is determined in step S404 that all keys have been processed, the process branches to step S407.
In step S407, the difference bit position table of the sort key smaller than the reference key is set in the difference bit position table, and the process proceeds to step S408 to extract the sort keys smaller than the reference key in the order of the difference bit position.

  Here, “the difference bit position table of the sort key smaller than the reference key is set in the difference bit position table in step S407” means that the difference bit position table to be processed in the next step S408 is smaller than the reference key. The difference bit position table of the sort key is used. In the example of FIG. 2B, the process using the difference bit position table 310a is performed. For example, it is executed by setting the address of the difference bit position table as the address of the difference bit position table of the sort key smaller than the reference key. The meaning of “set a difference bit position table for a sort key smaller than the reference key in the difference bit position table” described above is the same in the following other steps. Similarly, in the case of setting a difference bit position table of a larger sort key, the difference bit position table to be processed in the next step is to be a difference bit position table of a sort key larger than the reference key.

In step S408, “retrieve the sort key in the order of the difference bit position” means that the sort key is taken out in the ascending or descending order of the difference bit position with respect to the reference key, and for a set of a plurality of sort keys having the same difference bit position, The extraction of the sort keys in ascending or descending order of the difference bit positions with respect to the reference key in the set is repeated until all the keys are separated and extracted by the difference bit positions.
Details of the processing in step S408 will be described later with reference to FIGS.

Subsequent to step S408, in step 409, the reference key obtained in step S402 is set in the sorted key table.
Subsequently, in step S410, the difference bit position table of the sort key larger than the reference key is set in the difference bit position table, the process proceeds to step S411, the sort keys larger than the reference key are extracted in the order of the difference bit position, and the process is terminated.
Details of the processing in step S411 will be described later with reference to FIGS.

Next, details of the processing of step S406, step S408, and step S411 shown in FIG. 4 will be described.
FIG. 5 explains in detail the process of step S406 shown in FIG. 4, and the process of classifying the sort key group smaller than the reference key and the larger sort key group according to the difference bit position in the embodiment of the present invention. It is a figure explaining the example of a flow.

First, in step S501, it is determined whether the sort key is larger than the reference key. The sort key is extracted from the key table in step S405 shown in FIG.
If the sort key is larger than the reference key, the process proceeds to step S502. If the sort key is smaller than the reference key, the process proceeds to step S504. By the processing in step S501, the sort keys are grouped into a group smaller than the reference key and a group larger than the reference key. Then, a sort key classification process is performed for each group.

  In step S502, the difference bit position table of the sort key larger than the reference key is set in the difference bit position table. Then, the process proceeds to step S503, where the sort key is classified based on the difference bit position between the sort key larger than the reference key and the reference key, and the process ends. Details of the processing in step S503 will be described later with reference to FIGS. 6A and 6B.

  On the other hand, if the process proceeds to step S504, the difference bit position table of the sort key smaller than the reference key is set in the difference bit position table. Then, the process proceeds to step S505, where the sort key is classified based on the difference bit position between the sort key smaller than the reference key and the reference key, and the process ends. Details of the processing in step S505 will be described later with reference to FIGS. 7A and 7B.

Next, with reference to FIG. 6A and FIG. 6B, processing for classifying a group of sort keys larger than the reference key in one embodiment of the present invention will be described. 6A and 6B are detailed process flow examples of the process in step S503 shown in FIG. 5 and the process in step S1103 shown in FIG. 11 described later.
FIG. 6A is a diagram for explaining an example of the processing flow in the previous stage of the processing for classifying a group of sort keys larger than the reference key in one embodiment of the present invention.

As shown in FIG. 6A, in step S601, the key set as the sort key and the key set as the reference key are compared as a bit string, and the bit position of the first non-matching bit viewed from the upper 0 bit is represented as a difference bit. Set to position. The sort key is set in step S405 shown in FIG. 4 or step S1101 shown in FIG. 11 described later, and is a target key for the classification process shown in FIGS. 6A and 6B.

In step S602, the parent link pointed to by the difference bit position set in step S601 is read from the difference bit position table. In step S603, it is determined whether the parent link has been set. If the parent link has not been set, the process proceeds to step S604, where the sort key read position is set to the parent link pointed to by the difference bit position in the difference bit position table, and in step S605, the sort key read position is indicated. The sorting key reading position is set for the link, and the process is terminated.

  In the difference bit position table, as described above, the reading position of the largest key among the keys having the same difference bit position is stored as the parent link. However, in the process of step S604, the first sort key is stored. As the provisional maximum value key, the read position of the sort key is set to the parent link pointed to by the difference bit position in the difference bit position table. Once the parent link has been set, that is, when the determination in step S603 has been set, the process proceeds to the subsequent process shown in FIG. 6B, and the maximum value key is updated by comparing the sort key and the maximum value key. Finally, among the keys having the same differential bit position, the reading position of the largest key is stored in the differential bit position table as a parent link.

FIG. 6B is a diagram illustrating a processing flow example at the latter stage of the process of classifying a group of sort keys larger than the reference key in one embodiment of the present invention.
As shown in FIG. 6B, in step S606, the key pointed to by the parent link is read from the key table as a parent sort key. In step S607, it is determined whether the sort key is larger than the parent sort key.
If it is determined in step S607 that the sort key is larger than the parent sort key, the process advances to step S608 to set the read position of the sort key in the parent link pointed to by the difference bit position in the difference bit position table, and in step S609, the sort key of the link table. The parent link obtained in step S602 is set to the link pointed to by the reading position, and the process ends.

The processing in step S608 described above is performed in order to set the reading position of the largest key among the keys having the same difference bit position in the parent link pointed to by the difference bit position in the difference bit position table. To update the parent link. Then, the process of step S609 sets the parent link set in the difference bit position table to the link of the link table pointed to by the read position of the sort key newly set in the difference bit position table. The information for tracing the reading position in the key table of the key having the same differential bit position is held.

  On the other hand, if it is determined in step S607 that the sort key is not larger than the parent sort key, the process proceeds to step S610, and the link pointed to by the parent link is read from the link table as the next link. In step S611, the next link and the parent link are the same. judge.

  If it is determined in step S611 that the next link and the parent link are the same, in step S612, the read position of the sort key is set to the link pointed to by the parent link in the link table. In step S613, the sorting key reading position is set in the link indicated by the sorting key reading position in the link table, and the process ends.

  On the other hand, if it is determined in step S611 that the next link and the parent link are not the same, the process advances to step S614 to set the sort key reading position for the link indicated by the parent link in the link table. In step S615, the next link obtained in step S610 is set to the link pointed to by the reading position of the sort key in the link table, and the process ends.

  The determination in step S611 described above is the same as the determination as to whether there is only one classified key among keys having the same differential bit position. If there is only one classified key, the read position is stored in the link of the link table indicated by the read position of this key. Therefore, it is updated at the read position of the sort key, and the read position of the sort key is set in the link of the link table pointed to by the read position of the sort key to indicate that the sort key is the last classified key at that time. When there are two or more classified keys, the sorting key reading position is inserted between the link relationship between the parent link and the next link expressed in the link table by the processing of step S614 and step S615.

Next, with reference to FIGS. 7A and 7B, processing for classifying a group of sort keys smaller than the reference key in one embodiment of the present invention will be described. 7A and 7B are detailed process flow examples of the process in step S505 shown in FIG. 5 and the process in step S903 shown in FIG. 9 described later.

FIG. 7A and FIG. 7B regarding the classification of the sort key group smaller than the reference key and the description of FIG. 6A and FIG. 6B regarding the classification of the sort key group larger than the reference key are compared. The processing flow examples are the same. 7B and FIG. 6B also show a processing flow in the latter stage, in which the sort key is compared with the parent sort key in step S707 in FIG. 7B, and it is determined whether the sort key is smaller than the parent sort key, whereas in step S607 in FIG. The other steps are substantially the same, only differing by determining whether it is greater than the parent sort key.
Therefore, the description of the process of classifying a group of sort keys smaller than the reference key in the embodiment of the present invention with reference to FIGS. 7A and 7B is completed.

Next, details of the processing in step 408 shown in FIG. 4 will be described.
FIG. 8 is a diagram for explaining a processing flow example of processing for extracting a sort key smaller than the reference key in order of difference bit position in the embodiment of the present invention, and is a diagram for explaining details of processing in step 408 shown in FIG. is there.

  As shown in FIG. 8, in step S801, the difference bit position in the difference bit position table, which is the primary storage area (not shown), of the difference bit position in the difference bit position table as an initial value, that is, the difference bit position that the sort target key may have. Set the maximum value. In the example of FIG. 2B, 3 is set as the maximum value of the difference bit position.

  In step S802, it is determined whether entries for all the difference bit positions in the difference bit position table have been processed. If processed, the sort process is terminated, and if not processed, the process proceeds to step S803.

  In step S803, the parent link is read from the entry indicated by the difference bit position from the difference bit position table. If a parent link is set in the entry pointed to by the difference bit position, the read parent link is set as a parent link which is a primary storage area. That is, a notation such as “read parent link” in the description of this embodiment may mean that the parent link is read and set to a parent link which is a primary storage area (not shown). The same applies to items other than the parent link.

In the next step S804, it is determined whether the parent link read from the entry indicated by the difference bit position in step S803 has been set. If not set, the process branches to step S809, the value is decremented by 1 from the difference bit position, and the process returns to step S802. If set, the process proceeds to step S805. In the example of FIG. 2B, the parent link is not set in the differential bit position 3 that is the maximum value, and the parent links P2 and P1 are set in the differential bit position 2 and the differential bit position 0, respectively.

  In step S805, the key pointed to by the parent link (key table read position) set in step S803 is extracted from the key table as a sorted key, and the extracted sorted key is set in the sorted key table. As described above, since the key whose read position is stored in the difference bit position table is the maximum value key among the keys classified into the same difference bit position, the sorted key table in step S805. Set to. In the example of FIG. 2B, at the difference bit position 2, the key “1000” in the key table pointed to by the parent link P2 is extracted as a sorted key and set in the sorted key table. At the difference bit position 0, the key “0111” in the key table pointed to by the parent link P1 is extracted as a sorted key and set in the sorted key table.

  In step S806, the link pointed to by the parent link set in step S803 is read from the link table as the next link. In step S807, the parent link pointed to by the difference bit position in the difference bit position table is deleted, and the flow returns to step S808. move on. Here, the parent link pointed to by the difference bit position in the difference bit position table is deleted because the classification process of the maximum value key whose read position of the key table is the parent link has been completed in step S805, and is no longer in the difference bit position table. This is because the above parent link is not necessary and the influence on the subsequent processing due to this parent link remaining in the difference bit position table is avoided.

In step S808, it is determined whether the next link read in step S806 and the parent link set in step S803 are the same.
If it is determined in step S808 that the next link and the parent link are the same, the process returns to step S802 via step S809 described above.
By the loop processing from step S802 to step S809 described above, the sorted key is extracted and stored in the sorted key table in descending order of the difference bit position. Since the value of the key having the larger difference bit position is closer to the value of the reference key, the sorted key is extracted in descending order by extracting the sorted key in descending order of the difference bit position.

Therefore, the first sorted key table storing the sort key smaller than the reference key is provided separately from the second sorted key table storing the sort key larger than the reference key, and the one having the larger address of the first sorted key table is stored. By storing the sorted keys in descending order from the first to the smaller, the sorted keys can be read from the first sorted key table in ascending order.
Further, the number of sort keys smaller than the reference key and the number of sort keys larger than the reference key are counted according to the determination result of step S501 shown in FIG. 5, and the range and reference key for storing the sort key smaller than the reference key in the sorted key table are stored. It is also possible to determine the storage location and the range for storing the sort key larger than the reference key.

  Furthermore, in order to extract the sort keys smaller than the reference key in ascending order, a minimum value can be set for the parent link of the difference bit position table, and the sort keys can be extracted as sorted keys in ascending order of the difference bit positions. However, when processing is performed in ascending order of the difference bit positions, there are a plurality of keys having the same difference bit position, and if the sub-classification by the difference bit position described later is performed, the data of the original difference bit position table is destroyed. Since there is a possibility, it is necessary to prepare a differential bit position table for fine classification separately.

  In step S808 described above, it is determined that the next link and the parent link are the same when the key having the difference bit position where the parent link is stored has only a key having the parent link as the reading position of the key table. is there. In the example of FIG. 2B, the reading position 621 of the link table 311 is P2, and the link 622 is P2. Further, in the example of FIG. 2A, this corresponds to the case where the key group 142a is composed of only the key “1000” and the classification process for the key group 142a is not performed. In this case, this means that the classification process for the difference bit position for which the parent link is stored in the difference bit position table has been completed. Therefore, the value 1 is reduced from the difference bit position, and step S802 for the next difference bit position is performed. The process proceeds to S808. In the example of FIG. 2A, the process for the key group 140a at the difference bit position 0 is performed.

  On the other hand, if it is determined in step S808 that the next link and the parent link are not the same, the process proceeds to step S810, and the sorted key is set as the reference key. The sorted key is taken out in the above-described step S805 and set in the sorted key table. In order to use this sorted key as a reference key in sub-classification described later, a reference key which is a temporary storage area (not shown) Set to. In step S811, a plurality of keys having the same difference bit position (set in step S801 or step S809) are further classified (subclassified) by the lower-order difference bit positions. Details of the processing in step S811 will be described later with reference to FIG.

When it is determined in step S808 that the next link and the parent link are not the same, in the example of FIG. 2B, the reading position 621 of the link table 311 is P1 and the link 622 is P4. In the example of FIG. 2A, classification 149b based on difference bit positions is executed for a plurality of keys in the key group 140a.
In the process of step S811, the difference bit position table and the link table are reset, and the process returns to step S801. Then, the process for the set differential bit position of the parent link in the reset differential bit position table is repeated.

  The above-described processing is repeated until it is determined in step S802 that processing has been completed for all the differential bit position entries in the differential bit position table. When the determination is obtained, sorting of the sorting target keys is completed. The process is terminated.

Next, with reference to FIG. 9, the process of step S811 shown in FIG. 8 will be described in detail.
FIG. 9 is a diagram for explaining a detailed processing flow example of processing for further classifying a plurality of keys having the same differential bit position in the embodiment of the present invention.

  As shown in FIG. 9, in step S901, the key pointed to by the next link is extracted from the key table as a sort key, and the next link is set as the sort key reading position. In step S902, the link pointed to by the next link is read from the link table and saved in the save area. The next link in step S901 is a link indicated by the parent link read from the link table in step S806 shown in FIG. In the case of the classification processing of the key having the difference bit position 0 in the example of FIG. 2B, the parent link is P1, the next link is P4, and the link pointed to by the next link and saved in the save area is P3.

  The reason for saving the link pointed to by the next link in step S902 is that in the next step S903, the sort key based on the difference bit position of the sort key extracted in step S901 with respect to the reference key set in step S811 shown in FIG. 8 is performed. As a result, since the value of the link pointed to by the next link in the link table is rewritten, it is saved in advance and the processing using the link pointed to by the next link after step S903 is normally performed.

In step S903, the process of obtaining the difference bit position between the sort key and the reference key and classifying the sort key based on the difference bit position, which has been described in detail with reference to FIGS. 7A and 7B, is executed.
In step S904, it is determined whether the next link and the save link match. If the evacuation link matches the next link, the process ends because there are no remaining keys to be classified having the same differential bit position.

  On the other hand, if it is determined in step S904 that the save link does not match the next link, the save link is set as the next link in step S905, and the process returns to step S901. The newly set next link is set as the key table read position. The classification process is continued using the key to be sorted as the sort key to be classified.

  The loop processing in steps S901 to S905 described above is repeated until the save link matches the next link, that is, until there is no key to be classified, and the processing is ended when there is no key to be classified.

Next, details of the processing in step 411 shown in FIG. 4 will be described.
FIG. 10 is a diagram for explaining a processing flow example of processing for extracting a sort key larger than the reference key in the order of difference bit positions according to the embodiment of the present invention, and is a diagram for explaining details of processing in step 411 shown in FIG. is there.

  Compared with the example of the processing flow for extracting the sort key smaller than the reference key shown in FIG. 8 in the order of the difference bit positions, the example of the processing flow for extracting the sort key larger than the reference key shown in FIG. The details of the process of further classifying a plurality of keys having the same differential bit position are the same as those shown in FIG. 8 except that the details are as shown in FIG. Therefore, further description of the processing flow example shown in FIG. 10 is omitted.

FIG. 11 shows details of the processing in step S1103 shown in FIG. 10, and a detailed processing flow example of processing for further classifying a plurality of keys having the same differential bit position in one embodiment of the present invention. It is a figure explaining.
A detailed processing flow example of the process of further classifying a plurality of keys having the same difference bit position shown in FIG. 9 and FIG. 11 shows a process for further classifying a plurality of keys having the same difference bit position. While the process of step S903 shown in FIG. 9 classifies the sort key based on the difference bit position between the sort key and the reference key smaller than the reference key, the process of step S1103 shown in FIG. The sort keys are the same except that the sort keys are classified according to the difference bit positions.
The details of the process in step S903 shown in FIG. 9 are as shown in FIGS. 7A and 7B, whereas the details of the process in step S1103 shown in FIG. 11 are as shown in FIGS. 6A and 6B. Therefore, further description of the processing flow example shown in FIG. 11 is omitted.

  FIG. 12 is a diagram for explaining the flow of data for extracting keys in descending order according to an embodiment of the present invention. As an example shown in FIG. 12, the same key as illustrated in FIGS. 2A and 2B is classified into a group of keys smaller than the reference key after classification into a group of keys larger than the reference key and smaller keys. This is a data flow for extracting keys in descending order.

FIG. 12A shows the initial value shown in FIG. 2B in the difference bit position table 310 a and the link table 311, and the first sorted key “1000” is extracted and stored in the sorted key table 313. It is the state that was done.
The difference bit position table 310a is accessed in descending order of the difference bit positions as in the example of the processing flow shown in FIG.

First, the parent link P2 stored in the entry whose differential bit position 611a indicated by the arrow 682a is 2 is read, and as indicated by the arrow 672b, the reading position 621 of the link table 311 is the link 622 from the entry of P2. Is read out as the next link. Since it is determined that both the parent link and the next link are the same at P2, the key “1000” is extracted as the sorted key from the reading position P2 of the key table 309 indicated by the arrow 682b, and the sorted key table 313 Stored in entry 640d.

  Next, the parent link P1 stored in the entry in which the differential bit position 611a indicated by the arrow 680a is 0 is read, and the reading position 621 in the link table 311 is the link 622 from the entry of P1 as indicated by the arrow 671b. P4 is read as the next link, and it is determined that the parent link P1 and the next link P4 are not the same.

  FIG. 12B shows the reference key “0111” in the sub-classification of the group of keys at the difference bit position 0 with respect to the reference key “1010”, and stores it in the sorted key table 313. The difference bit position table and the link table are reset by the reference key in the fine classification.

  The key “0111” is read from the read position P1 of the key table 309 indicated by the arrow 681b indicated by the parent link P1 stored in the entry in which the difference bit position 611a of the difference bit position table 310a shown in FIG. As shown by a dotted arrow 681d, it is taken out into the sorted key 650b and stored in the entry 640c of the sorted key table 313 as shown by an arrow 681c.

  On the other hand, based on the link relationship indicated by arrows 674c and 673c in the link table 311 shown in FIG. 12A, the keys at the reading positions P3 and P4 are read from the key table 309, and the sorted key 650b is used as the reference key. As a result, the difference bit position is recalculated, and the difference bit position table 310a and the link table 311 are reset.

  As indicated by the dotted arrow 661a, the difference bit position of the key “0010” stored at the read position P3 of the key table 309 is 1, so that the difference bit position 611a of the difference bit position table 310a is 1. Is set to P3. Similarly, as indicated by the dotted arrow 663a, the difference bit position of the key “0110” stored at the read position P4 of the key table 309 is 3, so that the difference bit position 611a of the difference bit position table 310a is 3 Is set to P4.

  In addition, since there is one key having a difference bit position of 1 and one key having a difference bit position of 3, each read position 621 of the link table 311 is P3 and P4 as indicated by arrows 673b and 674b. P3 and P4 are set in the entry link 622, respectively.

FIG. 12C shows that the keys stored in the entry whose reading positions 601 of the key table 309 are P3 and P4 are stored in the sorted key table 313, and the group of keys smaller than the reference key is sorted. It is in a completed state.

First, the parent link P4 stored in the entry whose difference bit position 611a indicated by the arrow 683a is 3 is read, and as indicated by the arrow 674b, the reading position 621 of the link table 311 is the link 622 from the entry of P4. Is read out as the next link. Since it is determined that both the parent link and the next link are the same at P4, the key “0110” is extracted as the sorted key from the reading position P4 of the key table 309 indicated by the arrow 684b, and the sorted key table 313 Stored in entry 640b.

Next, the parent link P3 stored in the entry whose difference bit position 611a indicated by the arrow 681a is 1 is read, and the reading position 621 of the link table 311 is the link 622 from the entry of P3 as indicated by the arrow 673b. P3 is read as the next link. Since it is determined that both the parent link and the next link are the same at P3, the key “0010” is extracted as the sorted key from the reading position P3 of the key table 309 indicated by the arrow 683b, and the sorted key table 313 Stored in entry 640a.
Since the parent link is not set for the entry whose difference bit position 611a is 0, the sorting of the group of keys smaller than the reference key is completed.

Although the embodiment for carrying out the present invention has been described in detail above, it is obvious to those skilled in the art that the embodiment of the present invention is not limited thereto and can be variously modified.
In addition, it is obvious that the bit string data sorting apparatus of the present invention can be constructed on a computer by a storage unit that stores data having the data structure illustrated in FIG. 2B and a program that causes the computer to execute the processing illustrated in FIG. .
Therefore, the program and a computer-readable recording medium recording the program are included in the embodiment of the present invention.
By using the new bit string data sorting method provided by the present invention described in detail above, it is possible to sort bit string data at higher speed.

100 Key sequence 130 Sorted key sequence 149a, 149b Classification by difference bit position 200 Bit string data sort device 205 Grouping means 210 Difference bit position calculation means 220 Difference bit position classification means 230 Sorted key output means 240 Control means 301 Data processing apparatus 302 Central processing unit 303 Cache memory 304 Bus 305 Main storage unit 306 External storage unit 307 Communication unit 308 Data storage unit 309 Key table 310a, 310b Difference bit position table 311 Link table 313 Sorted key table

Claims (10)

In the bit string data sort device that performs the sort process of the key to be sorted represented by the bit string,
Sort target key storage means for storing the sort target key;
A group of keys smaller than the reference key and a group of keys larger than the reference key by comparing the reference key, which is an arbitrary key among the whole sort target keys, with a key other than the reference key stored in the sort target key storage unit Grouping means for grouping, and
The bit that is initially different by the bit string comparison between the key that is the classification target of the classification process that constitutes the sort process stored in the sort target key storage unit and the key that is the reference among the keys that are the classification target A difference bit position calculation means for obtaining a difference bit position as a value;
Classification target key difference bit position storage means for storing identification information of a key having the difference bit position for each difference bit position obtained by the difference bit position calculation means;
For each difference bit position obtained by the difference bit position calculation means, the identification information of the key having the difference bit position is written in the classification target key difference bit position storage means, so that the key to be classified is the same difference. Differential bit position classification means for classifying into sets having bit positions;
When there is one key having the same differential bit position classified by the differential bit position classification means, that key is used, and when there are a plurality of keys having the same differential bit position, it becomes a reference in the next classification process. Sorted key output means for outputting keys as sorted keys,
Control means for controlling the grouping means, the difference bit position calculation means, the difference bit position classification means and the sorted key output means;
With
The control means includes
In the first classification process,
The grouping means performs grouping of a group of keys smaller than the reference key and a group of larger keys by comparing the size of the reference key and a key other than the reference key;
The difference bit position calculation means obtains a difference bit position between the reference key and a key other than the reference key,
The difference bit position classification means includes identification information of a key having the difference bit position for each of the difference bit positions obtained by the difference bit position calculation means for each group of keys smaller and larger than the reference key. Write to the classification target key difference bit position storage means,
In the subsequent classification process, for each group of keys smaller and larger than the reference key,
The difference bit position calculation means repeatedly obtains the difference bit position as a classification target of a set of keys including a plurality of keys having the same difference bit position classified by the difference bit position classification means,
The difference bit position classification means repeats classifying into a set having the same difference bit position for each difference bit position obtained by the difference bit position calculation means,
Each time the sorted key output means classifies the difference bit position classification means into a set having the same difference bit position for each difference bit position, the key in the case where there is one key having the same difference bit position And to output the reference key as a sorted key,
Controlling the grouping means, the difference bit position calculation means, the difference bit position classification means and the sorted key output means;
A bit string data sorting device characterized by the above. The bit string data sorting device according to claim 1,
The reference key is a key that is first read from the sort target key storage unit, and a key that is a reference among the keys to be classified in a group of keys smaller than the reference key is a maximum value key that takes a maximum value. A bit string data sorting device characterized in that a key serving as a reference among keys to be classified in a group of keys larger than the reference key is a minimum value key having a minimum value. In the bit string data sorting device according to claim 2,
The information for identifying the key is a reading position that is an address of the key in the sort target key storage unit in which the key is stored,
The classification target key difference bit position storage means includes the key to be classified in the group of keys smaller than the reference key and the maximum value key or the reference key that is a reference key among the keys to be classified. A first difference bit position table storing a parent link that is the reading position of the largest key among the keys having the same difference bit position with respect to the maximum value key or the reference key, for each difference bit position; The minimum value key for each difference bit position between the key to be classified in the group of keys larger than the reference key and the minimum value key to be a reference key among the keys to be classified or the reference key Alternatively, the parent link that is the reading position of the smallest key among the keys having the same differential bit position with respect to the reference key is stored. A bit string comprising a second difference bit position table and a link table storing a link which is a read position of the key having the same difference bit position as the key of the read position for each of the read positions Data sort device. In the bit string data sorting device according to claim 3,
The link having the read position of the sort target key storage means of the key having the same differential bit position as the read position of the link table is the link as the read position of the link table, and the link set to the read position is It is set to refer to the read position of the sort target key storage means of all the keys having the same differential bit position by sequentially tracing the last of the keys having the same differential bit position. The bit string data sorting apparatus, wherein the key read position link stores the last key read position. In the bit string data sorting method for sorting the sort target key represented by the bit string,
A sort target key storage step of storing the sort target key in a sort target key storage means for storing the sort target key;
A group of keys smaller than the reference key and a group of keys larger than the reference key by comparing the reference key, which is an arbitrary key among the whole sort target keys, with a key other than the reference key stored in the sort target key storage unit A grouping step for grouping,
The bit that is initially different by the bit string comparison between the key that is the classification target of the classification process that constitutes the sort process stored in the sort target key storage unit and the key that is the reference among the keys that are the classification target A difference bit position calculation step for obtaining a difference bit position as a value;
For each difference bit position obtained in the difference bit position calculation step, the identification target key difference bit position storage means writes the identification information of the key having the difference bit position to make the key to be classified the same difference bit. A differential bit position classification step for classifying into sets having positions;
When there is one key having the same differential bit position classified in the differential bit position classification step, that key is used, and when there are a plurality of keys having the same differential bit position, it becomes a reference in the next classification process. A sorted key output step for outputting the key as a sorted key;
A control step for controlling the grouping step, the difference bit position calculation step, the difference bit position classification step, and the sorted key output step;
With
The control step includes
In the first classification process,
The grouping step performs grouping of a group of keys smaller than the reference key and a group of larger keys by comparing the size of the reference key and a key other than the reference key;
The difference bit position calculation step obtains a difference bit position between the reference key and a key other than the reference key,
In the differential bit position classification step, for each group of keys smaller and larger than the reference key, identification information of a key having the differential bit position for each differential bit position obtained in the differential bit position calculation step Write to the classification target key difference bit position storage means,
In the subsequent classification process, for each group of keys smaller and larger than the reference key,
The difference bit position calculation step repeats obtaining the difference bit position using a set of keys including a plurality of keys having the same difference bit position classified by the difference bit position classification means as the classification target,
The difference bit position classification step repeats classifying into a set having the same difference bit position for each difference bit position obtained by the difference bit position calculation means,
When the sorted key output step classifies the difference bit position classification step into a set having the same difference bit position for each difference bit position, the key in the case where there is one key having the same difference bit position And to output the reference key as a sorted key,
Controlling the grouping step, the difference bit position calculation step, the difference bit position classification step and the sorted key output step;
A bit string data sorting method characterized by the above. The bit string data sorting method according to claim 5,
The reference key is a key that is first read from the sort target key storage unit, and a key that is a reference among the keys to be classified in a group of keys smaller than the reference key is a maximum value key that takes a maximum value. A bit string data sorting method characterized in that a reference key among the keys to be classified in a group of keys larger than the reference key is a minimum value key having a minimum value. The bit string data sorting method according to claim 6,
The information for identifying the key is a reading position that is an address of the key in the sort target key storage unit in which the key is stored,
The classification target key difference bit position storage means includes the key to be classified in the group of keys smaller than the reference key and the maximum value key or the reference key that is a reference key among the keys to be classified. A first difference bit position table storing a parent link that is the reading position of the largest key among the keys having the same difference bit position with respect to the maximum value key or the reference key, for each difference bit position; The minimum value key for each difference bit position between the key to be classified in the group of keys larger than the reference key and the minimum value key to be a reference key among the keys to be classified or the reference key Alternatively, the parent link that is the reading position of the smallest key among the keys having the same differential bit position with respect to the reference key is stored. A bit string comprising a second difference bit position table and a link table storing a link which is a read position of the key having the same difference bit position as the key of the read position for each of the read positions Data sorting method. The bit string data sorting method according to claim 7,
The link having the read position of the sort target key storage means of the key having the same differential bit position as the read position of the link table is the link as the read position of the link table, and the link set to the read position is It is set to refer to the read position of the sort target key storage means of all the keys having the same differential bit position by sequentially tracing the last of the keys having the same differential bit position. The bit string data sorting method, wherein the key read position link stores the last key read position.   A program for causing a computer to execute the bit string data sorting method according to any one of claims 5 to 8. A computer-readable storage medium storing a program for causing a computer to execute the bit string data sorting method according to any one of claims 5 to 8.

Images