JPS6029989B2 - Korean sorting control system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a Korean language sorting control system, in particular, in ordering a Korean word string including Hangul characters, Chinese characters, and alphanumeric characters into an order corresponding to the reading. The present invention relates to a Korean language sorting control system which prepares a dictionary in which the pronunciations are numerically coded, indexes the dictionary, numerically codes the Korean words, and performs the above ordering.

In general, the above-mentioned Hangul characters can be thought of as phonetic characters composed of a combination of "consonant - vowel and one consonant", and are composed of a vowel group as shown in Figure 1A and a consonant group as shown in Figure 1B. configured. In addition, in Fig. 1 A and B,
Each upper row shows a Hangul glyme, and each lower row shows the corresponding pronunciation. Therefore, for example, the two characters shown in FIG. 2A are ``han■e'' shown in FIG.
It is pronounced as "l".

Recently, KEF (Korean
PrMessingE phantom ended Featme) code has been established.

However, since the KEF code is not uniquely associated with the pronunciation of each character, when ordering Korean words containing the above-mentioned Hangul characters according to their pronunciations, for example, to create a list, etc. The process becomes extremely complicated. The present invention aims to provide a system that solves this problem, and focuses on the fact that each Hangul character is composed of a combination of one or more graphemes, as shown in Figure 2. The object of this invention is to allocate a numerically encoded code corresponding to the reading of each grapheme, and to arrange the characters in ascending order and/or descending order based on the encoded numerical code.

Therefore, the Korean language sorting control system of the present invention selects word strings containing Hangul characters to which predetermined codes are assigned to Hangul characters, Chinese characters, and numbers. In a Korean sorting control system that uses codes ordered in an order corresponding to In addition to providing a character attribute dictionary that numerically encodes and stores the number of strokes in the word string, a sorting dictionary prepared from the character attribute dictionary or the character attribute dictionary for each character included in the word corresponds to each word in the word string. It is equipped with a numeric code generation unit that indexes a character attribute load module and generates a numeric code corresponding to the word, and each of the word strings in the word string is A sorting processing unit is provided for arranging words in ascending order and/or descending order of the numerical codes, and a corresponding word is extracted from each numerical code of a plurality of numerical code strings arranged by the sorting processing unit. It is characterized by what it did. This will be explained below with reference to the drawings. Fig. 3 is an overall configuration diagram of an embodiment of the system of the present invention, Figs. 4 A to E are explanatory diagrams explaining aspects of numerical codes used in the present invention, and Fig. 5 is stored in the character attribute dictionary shown in Fig. 3. 6 shows an example of the numeric code stored in the character attribute load module for sorting shown in FIG. 3, and FIG. 7 corresponds to a given word to be sorted. FIG. 8 is an embodiment of the character attribute dictionary, and FIG. 9 is an embodiment of the sorting character attribute load module. , Figures 10 A to D are explanatory diagrams explaining the initialization rules according to the present invention, Figures 11 A to C are explanatory diagrams explaining the handling of double consonants that exist in Korean, and Figure 12 is an explanatory diagram explaining the handling of double consonants. FIG. 13 is an explanatory diagram illustrating the sorting results obtained when handling correctly, and FIG. 13 is an explanatory diagram illustrating the sorting results when the initialization rule is not applied and when it is applied. Third
In the figure, 1 is one of the given words, 2 is a numerical code generator, 3 is a sort processing unit, 4 is data for sorting,
5 represents the sorting result, 6 represents the character attribute dictionary, and 7 represents the character attribute load module for sorting.

The embodiment shown in FIG. 3 will be explained in more detail later, but conceptually speaking, it can be considered that the ordering process is performed as follows.

'1} Assume that word x 1 is now input as a word to be sorted.

In this case, for each character "Riku", and 7..., KE
The character attribute dictionary 6 and/or the sorting character attribute load module 7 are indexed using the F code as a key. ■
Then, a reading instruction part given as a result of extracting the reading numerical codes "265602", "023400", etc. corresponding to each character "Riku", etc., and a double consonant indication part that indicates the presence of a double consonant, which will be described later. , generates a numerical code consisting of a stroke count instruction part that specifies the number of strokes for a kanji.

(31) The numerical code is connected to the head of the word XI to create sorting data 4.

'41 The sorting processing unit 3 considers the sorting data 4 to be a numerical value with weight, and orders the sorting data 4 in ascending or descending order, as is well known in the art.

{5} The sorting processing unit 3 deletes the numerical code given to the head of the sorting data of the ordered results and outputs the sorting result 5 of only words.

The manner in which words are numerically encoded will be described below. FIG. 4A shows a mode of numerically encoding the pronunciations of Hangul characters and Chinese characters. As mentioned above, Hangul characters are generally composed of a combination of three graphemes. Kanji readings can also be mapped to Hangul characters. These are called the first voice, middle voice, and final voice. When encoding the reading, one byte is used for the first voice, one byte for the middle voice, and one byte for the final voice, corresponding to a predetermined code.
A part-time job is prepared. If there is no final voice, ▼00▼ will be given with 16 ships displayed. FIG. 4B shows a mode in which so-called blank spaces are numerically encoded. In this case, 3 bytes are also prepared, but 1 byte is provided for the first byte.
▼00▼ is given in 6-Hayabusa display, and “
A KEF code ▼4040▼ indicating “blank” is given.
FIG. 4C shows a mode of converting alphanumeric characters into numerical codes. In this case as well, 3 bytes are prepared, the first 1 byte is given ▼FF▼ in hexadecimal notation, and the next 2 bytes are given KEF codes of alphanumeric characters.
In Korean, there are double consonants such as "Fufure" [[1, "d" and "Susu 1" shown in Figure 1 B. The handling of these double consonants will be explained later, but they can be For this reason, when the number of characters that make up one word is N, the number of bytes corresponding to the quotient of (N13)/4 is used as the double consonant indicator. Prepared as
Two bits are prepared for each character that makes up the word.

As shown in Figure 4D, when a double consonant is present in the first voice of a certain character, "1" is given to the first bit, and when a double consonant is present in the final voice of a certain character, "1" is given to the later bit. Given. Therefore, the 2 bits for a certain character contain "0".
One of "0", "10", "01", and "11" will be given. Furthermore, in the case of kanji, there are so-called homonyms when looking at the reading itself.

In consideration of this, a 1-byte stroke number indicating section is prepared for each character constituting the word to be sorted, as shown in FIG. 4E. In the case of a non-kanji character, ▼00▼ is given in a 16-hayabusa display, and in the case of a kanji character, the number of strokes of the kanji is given in a 16-hayabusa display. In the sorting data 4 shown in FIG. 3, for the input word XI, ■ the reading numerical value code "265602" corresponding to the Chinese character "Riku" is given, and the reading numerical value code "023400" corresponding to the Hangul character "Uchi" is given. is given, the numeric code ``004040'' is given for the blank space, and the numeric code ``1 for FFA'' is given for the alphabetic letter ``A'', which corresponds to the number ``1''. The numerical code "FFA cord 1" is given.

(B'Furthermore, in the double consonant indication section, the following numerical code r230000'' is given.

That is, since there are 9 characters in the input word x 1, (913)/4=3, and a 3-byte double consonant indicating part is prepared.

Since the character ``Riku'' does not have a double consonant, the 2 bits in Figure 4 D are ``roo'', and since the letter ``Wahu'' has a double consonant in the first voice, ``10'' is used, and the blank has a double consonant. ``00'' is given because the letter ``1'' has double consonants in the initial and final sounds, and ``11'' is given because the letter ``1'' has a double consonant in the first and final sounds. 0000,
...'', so ``230,000'' is given with 16 ships displayed. [C) Also, in the stroke number instruction section, the stroke numbers "Junichi" and "06" are given to the kanji characters "Riku" and "Meat" in the word X, respectively. FIG. 5 shows an embodiment of the numerical code stored in the character attribute dictionary shown in FIG.

As mentioned above, each Hangul character is composed of glymes. For this reason, numerical codes with an order corresponding to the pronunciation are assigned to each grapheme or grapheme group. That is, ▼02▼ is assigned to correspond to "7" shown in the figure, ▼03▼ is assigned to correspond to "77", and so on. Since the character attribute dictionary is not only used for sorting by reading, which is the object of the present invention, odd codes are given to double consonants as shown in the figure. On the other hand, FIG. 6 shows a similar numerical code stored in the sorting character attribute load module shown in FIG.

Since the load module is used only for sorting, the case differs from that shown in FIG. 5 in that the double consonant is given the same code as the original consonant. That is, the same code ▼02▼ is given to both the consonant "7" and the double consonant "77". The reason for this will be explained later. Regarding the character Rafu L in the sorting data shown in Figure 3, there is a numerical code ▼o2▼ corresponding to the double consonant “fufu”, a numerical code ▼milk▼ corresponding to the vowel “f”, and a final voice. One numeric code ▼023400▼ is given by the code ▼00▼ indicating that it is not possible.

Figure 7 shows, for example, that the words to be sorted are
This shows the format of sorting data that is generated when 4 characters (8 bytes) are given in the form of "key".

That is, for each of the four characters, there is a total of one three-byte numerical code as explained in connection with Figures 4A to C.
Two bytes are given. Also, as the double consonant designation part, (43)/4=1...Since the remainder is 3, 1 byte is given.

Furthermore, 4 bytes are given as a stroke number instruction section. Then, the word X into which the numerical code generated in this way is input
is added to the head of the file, and the data becomes sorting data 4 as a whole. The sorting data 4 is a weighted numerical code such that the leftmost digit in the figure is the most significant digit, and each sorting data is sorted by the sorting processing unit 3 illustrated in FIG. 3 based on the magnitude relationship of the numerical values. Kororashi・are arranged in descending order.
Then, from the result, only the word x portion is extracted, resulting in sorting result 5 shown in FIG. FIG. 8 shows an embodiment of the character attribute dictionary.

In the dictionary 6, a phonetic reading numerical value code and a reading numerical value code are given corresponding to each character. In the dictionary 6, double consonants are stored as having odd-valued codes. In the case of Hangul kanji, there is no distinction between Onyomi and Kunyomi. However, in Korean, as will be described later, there is an initialization rule, and the reading when the initialization rule is applied is stored as a numerical code as the illustrated numerical reading code. Of course, for characters to which the initial sound rule is not applied, the same numerical code is stored in the illustrated phonetic reading numerical code castle and the reading numerical code castle. To this end, as will be described later, when the character appears at the beginning of a word, the illustrated reading numerical code can be extracted as the code for the character, regardless of whether or not the initialization rule is applied. . FIG. 9 shows an embodiment of the sorting character attribute load module.

In the module 7, corresponding to each character, "total number of strokes information", "sound reading numerical code" and "double consonant information" are stored as shown in the figure.
, ``Numerical reading code (Kannon rule applied)'', etc. are stored. The total number of strokes information also exists in the character attribute dictionary 6 shown in FIG. 8, and is the information described with reference to FIG. 4E.
Furthermore, the on-yomi numerical code and the pronunciation value code have been explained in connection with the character attribute dictionary 6 shown in FIG. It is stored as a numeric code. However, as explained in connection with FIG. 6, the numerical code for the reading in the module 7 is rewritten to an even-numbered code for the co-digital consonant.
In order to compensate for this, the module 7 prepares one byte of "duplex consonant information" for each character. That is, as shown in the figure, depending on the state in which double consonants are present, "0○00○○00", "0000001○", "0
Either pattern ``○○○0001'' or ``00000011'' is stored as double consonant information.

The characters stored in the character attribute load module 7 for sorting shown in FIG.
The characters stored in the illustrated character attribute dictionary 6 are called second-level characters.

In the numerical code generation processing unit 2 shown in FIG. If so, the value of the code is subtracted by 1, and the above-mentioned bit is set to "1" as the double consonant information. The reason for this will be explained later with reference to FIG. The above-mentioned initialization rule will now be explained with reference to FIG. 10.

In Korean, when the Hangul characters shown in the upper rows of FIGS. 10A, B, and C appear at the beginning of a word, the Hangul characters shown in the lower rows of the diagram are used. That is, the pronunciation changes when it appears at the beginning of a word. For example, as shown in Figure 10D, the word "conscience"
When pronouncing , even though the Hangul character o originally corresponds as shown in the "Incorrect" column in the illustration, it is changed as shown in the "Shoichi" column in the illustration.From this, It can be considered that the numerical reading code ``joko'' shown in FIGS. 8 and 9 stores the numerical code corresponding to the character ``tsubasahi'' in correspondence with the character ``on''.

Then, the numerical code to be extracted is selected depending on whether or not the initial sound rule is applied. FIGS. 11A to 11C are explanatory diagrams illustrating how double consonants are handled in the sorting process.

When sorting the characters ``r'', ``waff L'', and ``fu!'', they should be arranged as shown in FIG. 11B. However, for these characters, character attribute dictionary 6
As is clear from Figure 5, for the letter "o" ▼
o23400▼ is given, and for the letter Waf L, ▼o3
3400▼ is given, and ▼02 ball 00▼ is given for the letter "Fl". As a result, if we compare and sort only the numerical codes corresponding to the readings, we can see that 02 spots 00ku0
Since 25CO minus 033400, Figure 11C
They will be arranged as shown.

This point should be improved by preparing double consonant information as sorting data 4 and using the same numerical code for double consonants as that for normal consonants.

That is, corresponding to the character "futo" ▼023400,...
・、Give 0 (heavy consonant range) ▼ and deal with the characters with the wrong character ▼02
3400'... ...'2 (heavy consonant range) ▼ is given, and corresponding to the letter ``al'', ▼02 rin 00...0 (heavy consonant range) ▼ is given. By doing this, 023400...0<02 3400...2<
02 90...0, and it becomes possible to arrange them correctly as shown in FIG. 11B.

FIG. 12 explains the sorting results obtained when the double consonants are treated correctly as described above.

When a plurality of words shown in the left column of the figure are words to be sorted, they are correctly arranged as shown in the right column of the figure.
For this purpose, as explained in connection with FIGS. 8 and 9, when characters existing in the character attribute dictionary 6, that is, characters of the second level, are indexed, the numerical code corresponding to the reading is odd. When the code has a numerical value, it is used by subtracting the value of the code by 1 and setting bit "1" in the double consonant information field.

FIG. 12 shows the sorting results according to the present invention when the initialization rule is not applied and when it is applied.

When the words shown in the left column of the illustration are the words to be sorted, if the initial sound rule is not applied, they will be arranged as shown in the upper row of the right column of the illustration, but if the initial sound rule is applied. are arranged as shown in the lower row of the right column in the figure. As explained above, according to the present invention, it is possible to perform a process of arranging Korean words in a form corresponding to their pronunciation.

By preparing the first level characters in a character attribute load module for sorting that can be accessed at high speed, it becomes possible to generate numerical codes corresponding to readings at high speed. Furthermore, for each Hangul character, we have prepared a numerical code that corresponds to the pronunciation when the initialization rule is applied, and a numerical code that corresponds to the reading when the initialization rule is not applied, so even when the initialization rule is applied, the characters can be arranged correctly. becomes possible.
Even for Hangul characters that are not subject to the initialization rule, the codes are stored in the above-mentioned pronunciation value code castle, so when sorting by applying the initialization rule, if the character is at the beginning of a word, Control can be simplified by extracting the value of the above-mentioned reading numerical code for characters.

[Brief explanation of the drawing]

Figures 1 and 2 are explanatory diagrams explaining the prerequisite problems of the present invention, Figure 3 is an overall configuration diagram of an embodiment of the system of the present invention, and Figures 4 A to E show aspects of the numerical code used in the present invention. An explanatory diagram to explain, FIG. 5 is an embodiment of the numerical code stored in the character attribute dictionary shown in FIG. 3, and FIG. 6 is a numerical code stored in the character attribute load module for sorting shown in FIG. 3. One embodiment of the character attribute dictionary, FIG. 7 is an explanatory diagram illustrating sorting data in which a numerical code is assigned corresponding to a given word to be sorted, and FIG. 8 is an embodiment of the above-mentioned character attribute dictionary. FIG. 9 shows an embodiment of the character attribute load module for sorting, FIGS. 10A to D are explanatory diagrams explaining the initialization rules of the present invention, and FIGS. 11A to C exist in Korean. Figure 12 is an explanatory diagram explaining the handling of double consonants, Figure 12 is an explanatory diagram explaining the sorting results obtained when handling double consonants correctly, Figure 13 is sorting when the initialization rule is not applied and when it is applied. An explanatory diagram explaining the results is shown. In the figure, 1 is a word, 2 is a numerical code generation section, 3 is a sort processing section, 4 is data for sorting, 5 is a sorting result, 6 is a character attribute dictionary, and 7 is a character attribute load module for sorting. Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 9 Figure 8 Figure 10 Figure 11 Figure 12 Figure 13

Claims (1)

[Claims] 1. Word strings containing Hangul characters to which predetermined codes are assigned to Hangul characters, Chinese characters, and alphanumeric characters, respectively, are ordered in an order corresponding to the readings of the Hangul characters and Chinese characters. In a Korean sorting control system that uses the given code to order, at least the reading of the character and the number of strokes given corresponding to each of the above Hangeul characters, Kanji, and alphanumeric characters are numerically coded. In addition to providing a character attribute dictionary for converting and storing the character attribute dictionary, the character attribute dictionary for sorting or a character attribute load module for sorting prepared from the character attribute dictionary is provided for each character included in the word corresponding to each word in the word string. A numeric code generation section that indexes and generates a numeric code corresponding to the word is provided, and based on the numeric code generated by the numeric code generation section, each word in the word string is arranged in ascending order of the numeric code. and/or a sorting processing section for arranging in descending order, and corresponding words are extracted from each numerical code of a plurality of numerical code strings as a result of the arrangement by the sorting processing section. Korean sorting control system. 2 The character attribute dictionary and/or the character attribute load module for sorting are provided with on-yomi numeric codes corresponding to the Kanji in the words, and readings when the initialization rule is not applied corresponding to the Hangul characters. The Korean language sorting control system according to claim 1, characterized in that the Korean language sorting control system has a numeric code corresponding to the ``pronunciation'' and a numeric code corresponding to the pronunciation when the initialization rule is applied.