JP2002297665A - Data processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To facilitate the management of data having a history. SOLUTION: When candidate character strings, etc., used for Kana (Japanese syllabary)-Kanji (Chinese character) conversion are managed by imparting them priority orders, each candidate character string (word) is registered in a hash table HTL to which a hash value HV and an order value OV are given and managed by using the hash table HTL. When the word is used, the position of data is obtained from the hash value and at the same time, the order value OV is replaced with a value larger the current maximum value.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data processing method and a kana-kanji conversion method to which the method is applied, and more particularly to a technique for handling a plurality of data together with a predetermined processing order.

[0002]

2. Description of the Related Art Conventionally, techniques for handling a plurality of data together with their processing order, for example, history, have been used in various fields of data processing. For example, in the kana-kanji conversion, when displaying a plurality of homophone candidates, candidate character strings are displayed in the order of use, and in order to realize this, the history of each candidate is managed. Things are going on.

[0003] If the data itself can be directly operated, such history management can be easily performed by creating a table by arranging the used data in order from the top. When the data in the table is used, the data is moved to the head and the remaining data is rearranged. When there is a limit on the size of the table, when new data is used beyond this limit, it is common to delete the oldest used data, which is the last data in the table.

[0004] On the other hand, when the data itself whose history is to be managed cannot be directly operated, for example, when stored in a plurality of kana-kanji conversion dictionaries, a table called a hash table is prepared. , Data is managed through this hash table. The hash table uses a hash function and specifies data through the table, and can operate the data at high speed. Taking Kana-Kanji conversion as an example, the position and the history of the corresponding conversion candidate can be known by using the reading as a key and drawing a hash table. In order to store the processing order (history) in such a method using a hash table, a method of fixing the position of data and attaching a pointer to the history or a method of fixing the position of data and There is known a method of assigning a value indicating a rank or the like to data.

[0005]

However, each of these methods has a problem that a large amount of time is required for searching and rearranging data. For example, in the case of operating the data itself as a table, it is necessary to perform a linear search of the data in order to sort the plurality of data in the history order, and the efficiency decreases as the number of data increases. However, in the individual method, the arrangement of the data itself is changed, so if a binary tree search or the like is introduced in order to facilitate the search, the search key must be changed every time the data is rearranged. It takes a lot of trouble.

On the other hand, in the management using a hash table or the like, if a method is employed in which the position of data is fixed and a pointer is provided in the history to manage the history, which of the two different data is If you try to find out whether it has been done, there is no other way than to perform a linear search until the two corresponding data are found. Even if two data are immediately found in the hash table,
This is because the table for managing the rank of two data is only provided with a pointer, and it is not possible to know the history of the two data without tracing the pointer. In addition, if a method of assigning a value indicating the rank to each data while fixing the position of the data is adopted, there is no need to perform a linear search, but every time a history is added, the rank value of the data whose rank has been changed is changed. Everything has to be rewritten, which again requires a lot of work.

[0007] The device of the present invention solves these problems,
It is an object of the present invention to easily operate a plurality of data and easily manage rank values indicating the rank of processing.

[0008]

A data processing method according to the present invention that solves at least a part of the above-mentioned problems is a data processing method that handles a plurality of data together with a predetermined processing order. Preparing a hash table corresponding to a plurality of data to be managed, and managing a rank value indicating a rank of at least a part of the plurality of data in accordance with a hash value referred to from the hash table. Then, when one of the data is specified, a hash value is obtained by referring to the hash table, and a ranking value corresponding to the hash value is updated using a relative magnitude relation.

An invention of a data processing apparatus corresponding to this data processing method is a data processing apparatus for handling a plurality of data together with a predetermined processing order, and a hash table corresponding to a plurality of data to be managed. Hash table storage means, and a management means for managing a rank value indicating a rank of at least a part of the plurality of data in accordance with a hash value referenced from the hash table,
Updating means for obtaining a hash value by referring to the hash table when one of the data is specified, and updating a ranking value corresponding to the hash value using a relative magnitude relation. It is a gist.

According to this data processing technique, a plurality of data are managed using a hash table. When one of the data is designated, a hash value is obtained by referring to the hash table, and the hash value is calculated. The ranking value corresponding to the value is updated using the relative magnitude relation. Therefore, the management of the rank values only requires a relative magnitude relationship, and even when data is designated, there is no need to rearrange the data accordingly. In addition, there is no need to perform special data processing such as changing pointers. Further, since the management of the rank value is updated based only on the relative magnitude relation, it is not necessary to perform a linear search to judge the magnitude relation between two or more data, thereby reducing the time required for processing. Can be.

Here, the order of the predetermined processing is as follows:
A history indicating the order in which the data was used can be considered. In data processing, history management is a process that is often encountered, and it is extremely useful to realize this by a simple method.

Updating of the rank value based on the relative magnitude relation is as follows:
At the start of the predetermined process, the rank value is set to the same value for a plurality of data, and the predetermined process is executed,
Each time the processing is performed, the rank value may be increased by monotonically increasing or monotonically decreasing. This technique can be realized by a very simple algorithm.

When the rank value monotonically increases or decreases, the value may gradually increase or decrease, possibly exceeding the range of values that can be handled. Therefore, the order value may be changed before the order value that is monotonously increased or decreased reaches a predetermined limit value. In this way, even if the rank value is monotonously increased or decreased, the data processing does not get stuck. As the initial value, for example, a value of 0 or a value in which all bits are 1 in a binary representation can be considered.

Such a change of the rank value can be easily realized by a process of increasing or decreasing the rank value already given by a predetermined value. For example, a ranking value of 8
If it is assumed that the value is handled in bits, the initial value is set to a value of 0, and the value is monotonically increased by an increment of 1, a process of subtracting 200 from all rank values may be performed before performing the process 255 times. In this case, the rank value is reduced uniformly, but the magnitude relationship between them is maintained. Of course, for some data, the ranking value may be negative. Although it may be treated as a minus, if it becomes a minus, it is also convenient to treat it as a value 0. That is, if the result exceeds or exceeds a predetermined range when the value increases or decreases by a predetermined value, data exceeding the predetermined range is set to a boundary value corresponding to the predetermined range. In this case, since the data exceeding the predetermined range is set to the boundary value corresponding to the predetermined range, a part of the rank according to the rank value is lost. It can be used like an array of candidate character strings in kana-kanji conversion.

In the above data processing method, at least a part of the plurality of data is arranged in a predetermined order, and a connection before and after the data is determined according to a magnitude relationship between the assigned rank values. A pointer is provided, and the history of data can be managed by changing the pointer. By providing a pointer in this way, it is possible to easily access data before and after the rank value.

Further, in this method, the pointer may be handled as follows. That is, when one of the already used data of at least a part of the plurality of data is used, the pointer indicating the connection before and after the data is traced, and the pointers of the data before and after are connected. And rewrite the pointer of the used data to a value indicating the head of the plurality of data,
The order value of the used data is rewritten to a value that monotonically increases or decreases from the value of the data located at the head. By performing such processing, it is possible to easily perform ranking when data that has already been used is used again.

When the data is used for the first time, it may be handled as follows. That is, when one of at least some of the plurality of data is used for the first time, the oldest used data is deleted, and the pointer of the data is traced to terminate the latest data. In place of the data, the data used for the first time is registered, and the pointer of the data used for the first time is rewritten to a value indicating the head of at least a part of the plurality of data, The order value of the used data is rewritten to a value that monotonically increases or decreases from the value of the data located at the head. In this way, even if the data is used for the first time, the order of use can be easily managed.

The present invention can be understood as a process of comparing data processed by the various methods described above. That is, the data comparison method of the present invention is a comparison method for comparing the order of processing of a plurality of data. At least two of the plurality of data processed by the above-described data processing method are ranked. The gist is to acquire values and compare the acquired rank values to determine the rank of data processing.

These inventions are all described as an invention of an apparatus for executing the above method, an invention of a program which is executed on a computer to realize the above functions, and an invention as a recording medium on which such a program is recorded. Each can be grasped. The device may realize the storage of the above-mentioned hash table, the management of the rank value, the update of the rank value, and the like by executing the program on the computer, or the device may be realized by a discrete circuit configuration. It may be. The program may be a known programming language such as C language, Pascal, Fortran, Kobol, or BASIC, and may be an object-oriented programming language or a language such as JavaScript. Recording media include flexible disk, CD-ROM, DVD-
ROM, semiconductor memory (ROM, PROM, EEPROM)
M, a flash memory, etc.). Of course, it is also possible to store these programs in a server placed on a network such as the Internet, and download and use the programs on a client computer.

[0020]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described below based on examples. (1) Apparatus Configuration in Embodiment: FIG. 1 is a schematic configuration diagram of an electronic organizer or a mobile type computer 20 that implements a data processing method according to an embodiment of the present invention. The computer includes a small keyboard 11 shown in FIG. 2 and a small liquid crystal display 29, and can input Japanese using the fifteen keys shown in FIG. As will be described later, Japanese input is performed through Kana-Kanji conversion of Romaji input, and candidate character strings and next candidates in Kana-Kanji conversion are all displayed on the liquid crystal display 29. The computer 20 has a function of connecting to the network 10 such as the Internet, and can exchange various information with another computer 40 in a network environment.

FIG. 1 shows the internal configuration of the computer 20. The computer 20 includes a network interface (NT-I / F) 21 for controlling data exchange with the network 10 via a modem or a router 18, a CPU 22 for processing, a ROM 23 for storing processing programs and fixed data, RAM2 as area
4. Timer 25 for managing time, liquid crystal display 29
A display circuit 26 for controlling the display on a computer, a hard disk (HD) 27 for storing a kana-kanji dictionary and the like and storing the processed data, and an input interface (I / F) 2 for controlling an interface with the keyboard 11 and the touch panel 12.
8 and so on. The touch panel 12 is transparent, is disposed so as to overlap the liquid crystal display 29, and functions as a pointing device. In this embodiment, a small electronic organizer or a mobile computer 20 is used as an example. However, the present invention is not limited to a small computer, and may be used in a desktop computer or a notebook computer. It is possible. In this case, the hard disk 27 need not be limited to a fixed type, but may be a removable type or a removable storage device (for example, a CD-ROM, a CD-R, a CD-RW, a DVD-R).
OM, DVD-RAM, flexible disk, etc.).

In this embodiment, data processing in a Japanese input processing program among various processing programs executed by the computer 20 will be described. Although the Japanese language processing program is stored in the ROM 23, it may be stored in the hard disk 27, expanded on the RAM 24 at the time of startup, and executed. Alternatively, the data may be read from the above-mentioned removable recording medium. Furthermore, the program may be read from another server via the network 10 and executed.

The keyboard 11 will be described with reference to FIG. The keyboard 11 has a total of 15 keys. The details of these keys include ten input keys 15a to 15j for inputting data composed of characters and symbols (hereinafter, referred to as character / symbol data), a "conversion" key in kana-kanji conversion in addition to a movement instruction. And two movement keys 16a and 16b used as a "next candidate" key, an enter key 17a for transferring the input character / symbol data to the application side, and an input for erasing the input character / symbol data Data cancel key 17b,
A help key 17c for displaying a method of inputting character / symbol data using the keyboard 11.

These keys are arranged in an area that can be operated by four fingers of one hand except the thumb. Also, input keys 15a to 15j and movement keys 16a, 16b
Are arranged in the same layout as the push dial of the telephone in four columns and three columns, and the input key 15a
15j are numbers from "0" to "9", and the movement keys 16a and 16b are "*" (asterisk) and "#".
The two symbols (sharp) are assigned in the same positional relationship as the push dial number of the telephone.

The input keys 15a to 15j include numerals from "0" to "9", Roman alphabet consisting of 19 alphabets, "," (comma), "." (Period), and "-" (hyphen). Are assigned, and the Roman character and the code are assigned to one input key in a multiplex manner.

In this embodiment, the input keys 15a to 15j
However, in order to be used almost equally in Japanese input, Roman letters and codes are assigned, and in order to realize such assignment, the frequency of use of each character and each code in data input is considered. I have. For example,
Input key 1 to which the most frequently used vowel "A" is assigned
For 5d, do not assign other Roman characters or codes redundantly,
The input key 15a to which the frequently used vowel "E" is assigned is repeatedly assigned a consonant "B" which is relatively infrequently used.

Whether one of numeric, Roman, and code data is input when one of the input keys 15a to 15j is operated is determined according to a set input mode. Here, the input mode refers to the character type of the data to be input. In the present embodiment, the input mode can be set to four types of characters: numerals, kana characters and kanji (hereinafter referred to as kana kanji), alphabetic characters and symbols. It is. For example, when the input key 15a is operated, the number "1" is input when the input mode is set to the numeric mode, and the Roman character "E" and the Roman character "B" are input when the input mode is set to the kana-kanji mode. I do. In the kana-kanji mode, a method of specifying a key intended by the input person from multiple assigned keys will be described later.

(2) Principle of Data Processing Method: In this embodiment, the data processing method of the present invention is applied to the management of the history of kana-kanji conversion candidates in Japanese input allowing ambiguity from the keyboard 11. However, for the sake of convenience, the basic principle will be described first with reference to FIG.
Then, an application example in Japanese input using the keyboard 11 will be described in detail.

FIG. 3 is an explanatory diagram showing a hash table used in data processing in this embodiment. In this embodiment, the Japanese input processing program executes option conversion called concept conversion. Concept conversion means that if you enter "Kudamono" and press the "Convert" key, the normal kana-kanji conversion program will display conversion candidates with the same reading, such as "fruit" and "tube". On the other hand, "strawberry"
It outputs candidates included in the input concept such as “pineapple” and “apple”. Such a concept transformation is
As illustrated in FIG. 4, the kana-kanji conversion dictionary DIC referred to by the Japanese input processing program of this embodiment is “read”
In addition to information such as “conversion candidates (kanji and katakana, etc.)” and “part of speech,” conceptual information such as “Kudamono”, “Yasai”, and “Animal” are also provided, which makes it possible. However, as shown in FIG. 4, the structure of the dictionary is arranged using “reading” as an index.
Is not directly associated with “ichigo”, “pineapple”, etc., and stored in the dictionary DIC. As shown in FIG. 3, a hash table HTL is provided for an input character string “Fruit”, and the position of each word in the dictionary DIC is obtained via this hash table. It is.

The hash table HTL is provided for each concept in the concept conversion, and basically has the hash value HV
And the ranking value OV. Initially, the order values OV of the hash table HTL provided corresponding to each concept conversion are all 0 in the initial stage. Figure 3 shows this situation.
(A). Therefore, the candidate character strings displayed by the concept conversion are displayed in a priority order according to the arrangement at the start of use. In the hash table HTL illustrated in FIG.
When the concept conversion is started by inputting “Fruit”, “Mikan” placed at the top of the hash table HTL is displayed as the first candidate character string, and when the “Next candidate” key is operated, the hash Next candidates are displayed in the order from the top of the table HTL, such as “apple, hana, pineapple,...”. At this time, the maximum value of the ranking value is 0,
This is stored as the maximum value MX.

From this state, for example, if “pineapple” is selected from the next candidates, as shown in FIG. 3B, the ranking value OV corresponding to the hash value 2 in the hash table HTL is: The value is set to a value 1 which is incremented by a value 1 with respect to a value 0 which is the maximum value MX of the ranking value up to then. At this time, the maximum value MX is also updated to the value 1.
Next, if the concept is converted again by "Fruit" and the word "Karin" of the hash value 46 is selected, similarly,
The ranking value corresponding to this "Karin" is set to the maximum value MX + 1, that is, the value 2. This state is shown in FIG.
Thereafter, each time a “fruit” is called by the concept conversion and a word is selected, the rank value of the selected word is set to the maximum value MX + 1 of the rank values set up to then.

In this way, words are successively called by the concept conversion, and when the maximum value MX exceeds the value 201, the value 100 is subtracted from all the rank values OV, and the process of re-assigning the rank value OV is performed. However, if the subtracted value becomes negative, the value is set to 0. This situation is shown in FIG.
(B). In this example, if the replacement is performed at the time of the kana-kanji conversion, extra time is required until the conversion is completed. Therefore, the computer 20 performs the replacement at a timing different from the conversion using a conversion waiting time or the like. Like that. For this reason, the replacement is performed at a timing when the maximum value MX exceeds 201. Since the order value OV is 8-bit data, the allowable range is 0 to 255. If the value does not exceed 255, the replacement timing can be changed at any time. Of course, if there is no delay in the processing, the processing may be performed when the maximum value MX exceeds the value 255.

Once the setting of the ranking value OV is started in this way, when concept conversion is started next, the word having the largest ranking value OV is displayed as the first candidate character string.
In the example shown in FIG. 3 (C), “Karin” having a rank value OV of 2 is the first candidate, and when the “Next candidate” key is operated, “Pineapple” with a rank value of 1 is displayed at the top. ,
In the following, candidate character strings are displayed in the order according to the hash table HTL. The history can be easily determined only by searching the column of the rank value of the hash table HTL. Note that in FIGS. 3 and 5, the hash values of the hash table HTL are set to values from 0 to 99, but this is an example, and can be varied depending on the number of conversion candidates.

(3) Data Processing Method in Japanese Input Processing (First Embodiment) Next, a data processing method performed in conjunction with the Japanese input processing in this embodiment will be described. In addition, the Japanese input can be connected phrase conversion,
For convenience of explanation, the following description will be made assuming that conversion is performed in word units. FIG. 6 is a flowchart showing a Japanese input processing routine. This process is started when Japanese is input using the keyboard 11, and it is first determined whether a key of the keyboard 11 has been pressed (step S100). When any key on the keyboard 11 is pressed, the type of the key is determined (step S1).
10) If the operated key is one of the input keys 15a to 15j, a process of adding the Roman character assigned to the key to the input character string is performed (step S120), and the process is repeated from the key input check again (step S120). S10
0).

If the operated key is the move key 16a (step S110), the conversion process is executed (step S130), assuming that it is a conversion instruction. As shown in detail in FIG. 7, the conversion process acquires a character string composed of Roman characters input so far (step S132), and refers to the created hash table H1 corresponding to the character string (step S1).
34), referring to the dictionary DIC using the hash value HV (step S136), displaying a candidate character string in a predetermined area on the liquid crystal display 29 (step S136).
138). At this time, keyboard 1
Since two or more Roman characters are assigned to each of the one input keys 15a to 15j, the character string is not uniquely determined. For example, if the input keys 15g, 15d, 15h, and 15d are sequentially operated in order to input a character string "Kasa", the obtained character string is "KASA"
"KAWA""NAKA""NAWA""KA,A"
There are six types, "NA, A". Among them, the character string including “,” is inappropriate from the viewpoint of the conversion rule from Roman characters to kana characters, and thus can be excluded from the character string for search. All four types of character strings that do not include "," are appropriate combinations as kana character strings for kana-kanji conversion.
These are all adopted, and the corresponding candidate character strings are read from the dictionary DIC via the hash table H1. Note that in the original processing, morphological analysis is performed before this to extract phrases, but for simplicity of explanation,
Here, since the idiom conversion is taken as an example, the description of processes such as morphological analysis, segmentation, and modification of segment length will be omitted.

FIG. 8 is an explanatory diagram showing an example of the hash table H1 in this case. In this example, "KASA""KA
With respect to the four character strings “WA”, “NAKA”, and “NAWA”, the next word is read from the dictionary DIC on the premise of idiom conversion. For "KAASA": "umbrella", "cap", "Kasa", "bulk", "bulk", "halo", "acne", For "KAWA": "river", "river", "side", "skin"
"Leather", "Kawa", "NASA": "Nasa""Nasa""None"
For "dead", "Nasa", and "NAWA": "rope", "Nawa", "seedling", "Nawa", "nawa". A hash table H1 is also formed for these words. When the conversion for this character string is activated for the first time, the hash table H1 contains, as shown in FIG.
The rank value OV contains the value 0.

After the character string is input, when the move key 16a for instructing the conversion is pressed and the conversion process is executed, the word having the highest rank value OV is converted using the rank value of the hash table H1. It is displayed on the liquid crystal display 29 as a candidate (step S130). Thereafter, the processing returns to key input and the above processing is repeated (step S10).
0). With the first conversion candidate displayed, the movement key 16b used as a "next candidate key" for requesting the next candidate
Is operated (step S110), as shown in FIG. 9, the hash table H1 is referred to (step S14).
2) The candidate having the next largest ranking value OV is stored in the dictionary D
It is obtained from the IC (step S144), and is displayed on the liquid crystal display 29 (step S146). At this time, the value of the ranking value OV is not rewritten yet, but the ranking value may be immediately rewritten every time the next candidate is displayed.

Thus, the movement key 16a for instructing the conversion
By operating the move key 16b for instructing the display of the next candidate or the next candidate, the candidate character strings are sequentially displayed, and after seeing any of the candidate character strings, the user can transfer the displayed character string to the application side. When the enter key 17a for instructing the transfer is operated, the operation of this key is determined (step S11).
0), as shown in FIG. 10, the candidate character string currently displayed as the candidate character string is written to the application program side (step S152). Next, a process of rewriting the ranking value OV of the hash table H1 corresponding to the determined candidate character string to a value larger by 1 than the maximum value MX is performed with reference to the value of the maximum value MX (step S154). S156). At this time, if the maximum value MX exceeds the value 250 (step S158), a process of changing the rank value OV of the hash table H1 is immediately executed (step S160). This process is a process of subtracting the value 150 from all the rank values OV of the hash table H1 and writing the value 0 when the result is negative. After executing this processing, the maximum value MX is updated to the maximum value in the hash table H1 (step S162). After the above processing,
Exiting to "NEXT", the Japanese input process is temporarily terminated. In the candidate character string shown in FIG.
"River" → "umbrella" → "river" → "rope" → "river" → "umbrella" →
FIG. 8B shows the state of the rank values after "kawa" is sequentially selected and determined. As shown, when the same word is discontinuously selected and confirmed, the ranking value is
In this case, the value becomes a discrete value, but in any case, the value is set as a monotonically increasing value.

According to the Japanese input processing described above, Japanese input can be easily performed using a limited number of keys 15a to 15j. In addition, at this time, a plurality of Roman character strings are obtained, and the histories in the plurality of readings can be managed although the dictionary DIC is referred to as a reading. In addition, the management of this history
It can be performed very easily, and there is no need to rearrange the data. Even if a new word is selected, there is no need to perform a process of rearranging the rank values over all data. For this reason, history management in Japanese input can be performed extremely easily and at high speed. In the above embodiment, idiom conversion has been described as an example. However, it is a matter of course that continuous phrase conversion involving morphological analysis can be similarly performed. In such a case, the kana character string is cut out from the input kana character string by a well-known method such as the multi-clause longest match method or the minimum cost method, and the candidate character strings for each of the phrases are similarly managed using a hash table. I just need.

(4) Data Processing Method in Japanese Input Processing (Second Embodiment) Next, another method of data processing in Japanese input processing will be described. In the Japanese input process according to the second embodiment, the outline of the process is as follows.
This is the same as the first embodiment, except for the method of managing the hash table H2. FIG. 11 is an explanatory diagram illustrating an example of the hash table H2 used in the second embodiment. In the second embodiment, the data management manages the history of the candidate character strings by using a hash table H2 including a hash value HV and a key KY, and a history data table DT associated via the key KY. In the example shown in FIG. 11, the input keys 15g, 15d, 15h, and 15d are sequentially operated as in the first embodiment, and “KASA”, “KAWA”, “NAKA”, and “NAW” are operated.
"A" is assumed to be a conversion target. As shown, the history data table DT combined with the hash table H2 via the key KY in the second embodiment
In addition to Y, the character string CH, and the rank value OV, it has pointers PP and PN indicating a rank connection between respective data. The pointer PP indicates the key KY of the character string CH of the immediately preceding rank value, and the pointer PN indicates the key KY of the character string CH of the next rank value. As shown in the figure, the key “KY” having the value “7” has the pointer PP
Contains a special value (value 255) indicating the end, so that it can be understood that this data is the lowest data of the ranking value OV. Similarly, in the data “river” whose key KY has a value of 8, the pointer 255 has a value of 255, which indicates that this data is the data having the highest rank value OV. The data selected at least once is registered in the history data table DT with a pointer, and becomes a target of ranking value management.

Using this data structure, data processing is performed as follows. [1] When a word registered in the history data table is selected: For example, when “River” registered in FIG. 11 is selected and determined, a hash table H2 is obtained from the hash value HV of this “River”. Can be used to determine the location of the data. Next, the pointers PP and PN of the history data table DT
, The data before and after the ranking value OV is known, and these pointers PP and PN are combined. That is, the corresponding data (in this example, the character string “kawa”) is extracted from the combination of the pointers. Then, the key KY of the determined data is inserted into the pointer PN of the data having the highest rank value (“river” in the example of FIG. 11), and the pointer PP of the determined data is set.
Is the key KY of the data with the highest ranking value
Insert As a result, the determined data is placed at the head of the sequence of data linked by the pointer in the order of the rank value. Then, a value larger than the rank value of the data having the highest rank value is put in the rank value of this data.
It is to be noted that, when the maximum value of the rank value exceeds a predetermined range, the replacement of the rank value is performed in the same manner as in the first embodiment. FIG. 12 shows the history data table DT after the replacement is completed with respect to the example of FIG. In the figure, the bold lines indicate the places where the numerical values have been replaced.

[2] Next, a case where a candidate character string not existing in the history data table DT is selected and determined will be described. For example, in FIG. 11, "umbrella" has not been used yet, so it is not listed in the history data table DT. At this time, if there is a space in the history data table, the newly used character string candidate is registered there, and if there is no space, the data with the smallest rank value (the data with the pointer PP indicating the end value 255). ). With the replacement, the pointer PP of the data indicated by the pointer PN is changed to a value 255 indicating the end, and the newly registered data is set as the data having the largest rank value.
The pointers PP and PN are replaced. That is, the key KY of the determined data (in this example, “umbrella”) is inserted into the pointer PN of the data with the highest rank value (“river” in the example of FIG. 11), and the pointer PP of the determined data is stored in the pointer PP of the determined data. , The key KY of the data having the highest ranking value. As a result, the newly used and determined data is placed at the head of the list of data combined by the pointer in the order of the rank value. Then, a value larger than the rank value of the data having the highest rank value is put in the rank value of this data. It is to be noted that, when the maximum value of the rank value exceeds a predetermined range, the replacement of the rank value is performed in the same manner as in the first embodiment. FIG. 13 shows the history data table DT after the pointers have been replaced by replacing “umbrella” with “sore” in the example of FIG. In the figure, the bold lines indicate the places where the numerical values have been replaced.

With the above processing, the same effects as in the first embodiment can be obtained. In particular, when it is necessary to compare the histories of two data, it is possible to know which data was used immediately before by only the relative magnitude relation between the rank values. Further, in this embodiment, since each data is linked by the pointer in the order of the rank value, a binary tree search or the like can be applied instead of the linear search, and the processing can be speeded up.

Although the embodiments of the present invention have been described above, the present invention is not limited to these embodiments at all, and can be implemented in various forms without departing from the gist of the present invention. Of course. For example, the rank value is set as a monotonically increasing value in the present embodiment, but may be managed as a monotonically decreasing value. Also, the increment is not limited to the value 1 but the value 2
It may be incremented by one. In addition, the data processing according to the present embodiment is used for managing the history of words having the same reading in the normal kana-kanji conversion, the history management other than the kana-kanji conversion, for example, the configuration used for managing the history of the database, or referred to on the network It may be applied to the management of the history of the accessed URL address or the history of the accessed file.
Further, in the present embodiment, various processes are realized by executing software by the CPU, but such processes may be realized by hardware. Furthermore, in the above-described embodiment, the keyboard 11 having a small number of keys is used. However, the data processing method of the present invention may be executed on a mobile phone or the like that performs Japanese input using such a keyboard 11. .

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram of a computer that executes a data processing method according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a configuration of a keyboard 11 used in the embodiment.

FIG. 3 is a hash table HTL for explaining the principle of the invention;
It is explanatory drawing which shows an example of.

FIG. 4 is an explanatory diagram illustrating the configuration of a dictionary DIC for kana-kanji conversion in the embodiment.

FIG. 5 is an explanatory diagram illustrating an example of how a hash table HTL is rewritten in concept conversion.

FIG. 6 is a flowchart showing a Japanese input processing routine in the first embodiment.

FIG. 7 is a flowchart illustrating details of conversion processing in the first embodiment.

FIG. 8 is an explanatory diagram illustrating an example of a hash table H1 in the first embodiment.

FIG. 9 is a flowchart illustrating details of a next candidate display process in the first embodiment.

FIG. 10 is a flowchart illustrating details of a determination process in the first embodiment.

FIG. 11 is an explanatory diagram illustrating an example of a hash table H2 and a history data table DT according to the second embodiment.

FIG. 12 is an explanatory diagram showing how the history data table DT is rewritten when one of the existing candidate character strings is selected and determined in the second embodiment.

FIG. 13 is an explanatory diagram showing how the history data table DT is rewritten when one of the new candidate character strings is selected and confirmed in the second embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 ... Network 11 ... Keyboard 12 ... Touch panel 15a-15j ... Input key 16a, 16b ... Move key 17a ... Enter key 17b ... Input data cancel key 17c ... Help key 18 ... Router 20 ... Computer 22 ... CPU 23 ... ROM 24 ... RAM 25 timer 26 display circuit 27 hard disk 29 liquid crystal display CH character string DT history data table H1 hash table H2 hash table HTL hash table KY key PN pointer PP pointer

Claims (15)

[Claims] 1. A data processing method for handling a plurality of data together with a predetermined processing order, wherein a hash table corresponding to a plurality of data to be managed is prepared, and a hash value referred to from the hash table is provided. Correspondingly, for at least a part of the plurality of data, a rank value indicating the rank is managed, and when one of the data is designated, a hash value is obtained by referring to the hash table, and the hash value is determined. A data processing method for updating a ranking value corresponding to a value using a relative magnitude relation. 2. The data processing method according to claim 1, wherein the order of the predetermined processing is a history indicating the order in which the data was used. 3. The data processing method according to claim 1, wherein the rank value is the same value for the plurality of data at the start of the predetermined process, and is executed in the predetermined process. Accordingly, a data processing method in which a value that monotonically changes in either increase or decrease is added each time processing is performed. 4. The data processing method according to claim 3, wherein the ranking value that is monotonically increased or decreased is replaced before the ranking value reaches a predetermined limit value. 5. The data processing method according to claim 4, wherein the change of the rank value is a process of increasing or decreasing a previously assigned rank value by a predetermined value. 6. The data processing method according to claim 5, wherein when the value is increased or decreased by the predetermined value, if the result exceeds a predetermined range, the data exceeding the predetermined range is converted to the predetermined range. Data processing method to set the boundary value corresponding to. 7. The data processing method according to claim 2, wherein the at least a part of the plurality of data is arranged in a predetermined order, and according to a magnitude relationship between the assigned rank values. A data processing method in which a pointer indicating a connection before and after the data is provided, and the history of the data is managed by changing the pointer. 8. The data processing method according to claim 7, wherein, when at least one of the plurality of data is used, one of the already used data is used before and after the data. Tracing a pointer indicating the connection of the data, changing the pointer so that the pointers of the preceding and following data are connected, rewriting the pointer of the used data to a value indicating the head of the plurality of data, A data processing method in which a rank value is rewritten to a value that monotonically increases or decreases from the value of the data located at the top. 9. The data processing method according to claim 7, wherein when one of at least a part of the plurality of data is used for the first time, the oldest used data is deleted. Along with the data pointer,
The latest data is set as the end, and the data used for the first time is registered in place of the data, and the pointer of the data used for the first time indicates the head of at least a part of the plurality of data. A method for processing data in which the rank value of the used data is rewritten to a value that monotonically increases or decreases from the value of the data located at the top. 10. A comparison method for comparing the order of processing of a plurality of data, wherein at least two of the plurality of data processed by the data processing method according to claim 1 are:
A data comparison method for acquiring the rank values and comparing the acquired rank values to determine the rank of data processing. 11. A data processing apparatus for handling a plurality of data together with a predetermined processing order, a hash table storing means for storing a hash table corresponding to the plurality of data to be managed, and referring to the hash table. Management means for managing a rank value indicating a rank of at least a part of the plurality of data in accordance with the hash value to be provided; and when one of the data is designated, refer to the hash table. A data processing device comprising: updating means for obtaining a hash value and updating a ranking value corresponding to the hash value. 12. A program for causing a computer to execute a data processing method for handling a plurality of data together with a predetermined processing order, comprising: a function of managing a hash table corresponding to a plurality of data to be managed; A function of managing a rank value indicating a rank of at least a part of the plurality of data in accordance with a hash value referred to from a table, and referencing the hash table when one of the data is designated And a function of causing a computer to obtain a hash value and update a rank value corresponding to the hash value. 13. A recording medium on which a program corresponding to a data processing method for handling a plurality of data together with a predetermined processing order is recorded, wherein a function of managing a hash table corresponding to the plurality of data to be managed is provided. A function of managing a rank value indicating a rank of at least a part of the plurality of data in accordance with a hash value referred to from the hash table; and, when one of the data is designated, the hash A recording medium in which a program that realizes a function of obtaining a hash value with reference to a table and updating a ranking value corresponding to the hash value is readable by a computer. 14. A kana-kanji conversion method for performing kana-kanji conversion with reference to a kana-kanji conversion dictionary using a character string obtained based on an array of key codes obtained by using a plurality of keys. Preparing a hash table corresponding to a plurality of conversion candidates obtained from the kana-kanji dictionary based on the key code array, and corresponding to a hash value referred to from the hash table; For at least a part, manage a ranking value indicating the history ranking, and when one of the conversion candidates is specified, obtain a hash value by referring to the hash table and calculate a ranking value corresponding to the hash value. Kana-Kanji conversion method to be updated. 15. The kana-kanji conversion method according to claim 14, wherein the key code array is an array of key codes obtained from a smaller number of keys than a minimum number of keys that uniquely designates a kana. A kana-kanji conversion method.