JPH0456350B2 - - Google Patents

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a Japanese information processing device, such as a word processor, that includes a kana-kanji conversion process, and in particular, it relates to a Japanese information processing device that includes a kana-kanji conversion process, such as a word processor, and in particular, a method for sorting/merging kanji according to their readings. This paper relates to the reading information retrieval method used in the event.

[Conventional technology]

Japanese information processing devices often output and display a list of candidate characters as a result of kana-kanji conversion, a list of in-system kanji, a list of additional kanji, and the like.

Kanji are encoded using JIS codes or JEF codes, but generally when outputting and displaying multiple kanji, the arrangement is coded so that the operator can easily recognize the desired kanji. This is usually done using kanji attributes such as reading, number of strokes, and radicals, rather than order.

Kanji attribute information is held in the form of a table, and the amount of information per character is, for example, 36 bytes for 6 readings (6 bytes for each type), 1 byte for the number of strokes, 1 byte for the radical, and so on. Reading accounts for 94% of the total.

FIG. 2 shows the structure of a kanji reading table. Each row of this reading table is composed of six entries #1 to #6, each registering the first to sixth readings of one kanji. The vertical arrangement of lines follows the order of kanji codes. If the number of kanji stored in this reading table is, for example, about 3000 characters, the size of the reading table is 6 (bytes) x 6 (entries) x 3000 (characters) = 108 Kbytes.

[Problem that the invention seeks to solve]

In conventional kanji reading tables, the length of the display area for the reading of each kanji is fixed; for example, in the above example, the length was 36 bytes for six types. However, even if a kanji has multiple readings, the number of readings is 2.
The problem was that more than half of the reading table was blank, reducing memory usage efficiency.

[Means for Solving the Problems] In order to solve the above problems, the present invention registers a reading table that collects only the readings of kanji and a pointer that indicates the corresponding reading in the reading table for each kanji. The table size is reduced by having a two-stage structure with a pointer table, and this structure allows Japanese information processing devices to read the readings of all kanji stored in the device in a fixed order. and a reading pointer table means having one or more reading pointers arranged to indicate positions in the reading table means representing all the readings of each kanji in the order of the kanji code. It is characterized in that the corresponding reading pointer in the reading pointer table means is read out based on the kanji code, and then the reading is read out from the corresponding position in the reading table means using the reading pointer.

[Action of the invention]

FIG. 3 shows the concept of the present invention. In the figure, 31 is a reading pointer table, 32 is a reading pointer entry string, 33 is a reading pointer entry, 34 is a reading table, and 35 is a reading entry.

The reading table 34 according to the present invention is obtained by extracting the readings of all kanji stored in a Japanese information processing device and sorting and integrating them, for example, in order of EBCDIC code, and each reading is separated from the kanji code of its source. They are organized and registered in a uniform manner. Therefore, river (kawa), skin (kawa)
All different kanji with the same reading, such as ``kawa'', have one reading. Since it is associated with the read entry 35, the amount of data can be compressed.

The reading pointer table 31 is for associating each kanji code with each reading entry of the reading table 34 described above.
A pointer corresponding to the kanji code, that is, a reading pointer, is registered for each kanji code. Each reading pointer entry column 32 corresponds to a separate kanji code,
Arranged in order of kanji code. For example, "river"
For ``Kawa'', ``Gawa'', and ``Sen'', the reading pointers are in the reading pointer entry column 3.
Registered during 2nd. The length of the reading pointer can be much shorter than the reading length.

In this way, by accessing the reading pointer entry string in the reading pointer table 31 that corresponds to the kanji code, the registered reading pointers are sequentially read out, and the reading table 34 is accessed using them, and the reading pointer entry string of the reading pointer table 31 corresponding to the kanji code is accessed. You can extract all the readings. Reading pointer table 31
The combination of the reading table 34 and the reading table 34 can be made smaller than the conventional reading table shown in FIG.

〔Example〕

The details of the present invention will be explained below with reference to Examples.

FIG. 1 shows the configuration of one embodiment of the present invention, where 1 is a Japanese language information processing device, 2 is a storage device, and 3 is a Japanese language information processing device.
is a table access control unit, 4 is a kanji code/address conversion unit, 5 is a reading pointer table pointer, 6 is a reading pointer entry selection unit, 7 is a reading table pointer, 8 is a reading pointer table,
9 is a reading pointer entry string, 10 is a reading pointer entry, 11 is a reading table, 12 is a reading entry, 13 is a kanji code input, 14 is a reading pointer output, and 15 is a reading output.

The table access control unit 3 controls access to the reading pointer table 8 and the reading table 11 for the kanji code input 13, and extracts all the readings registered for the kanji represented by the kanji code as the reading output 15.

The reading pointer table 8 has, for example, 3000 reading pointer entry columns 9 corresponding to 3000 types of characters (kanji codes). Each reading pointer entry column 9 has six reading pointer entries 10 in order to be able to register up to six different readings for each kanji, as in the case of the conventional reading table explained in FIG. There is. The length of each read pointer entry is composed of 2 bytes in this embodiment. Therefore, each read pointer entry string has a length of 12 bytes, and the entire read pointer table has a size of 36 Kbytes. The value of the reading pointer entry points directly to the reading entry in the reading table 11.

In this embodiment, the reading table 11 registers 3170 types of readings extracted from 3000 types of characters, and each reading entry 12 is arranged in order of EBCDIC code. Another reading entry 1
2 is 6 bytes long, and the entire size of the reading table 11 is about 20 Kbytes.

In operation, the table access control unit 3 instructs the kanji code/address conversion unit 4 to convert the address of the reading pointer table 8 corresponding to the kanji code, that is, the target reading from the top position of the table, in response to the kanji code input 13. The offset to the pointer entry column is calculated, the result is read, and the result is set in the pointer table pointer 5.

Next, the read pointer table 8 is accessed based on the value of the read pointer table pointer 5, and the read pointer output 14 is input to the read pointer entry selection section 6.

The reading pointer entry selection unit 6 selects one of the six reading pointer entries 10 included in the reading pointer entry column 9 from the reading pointer output 14 based on instructions from the table access control unit 3.
Select one and set its value, that is, the reading pointer, in the reading table pointer 7. However, if the reading pointer entry is empty, the next reading pointer entry is selected.

Next, the reading table 11 is accessed according to the reading pointer set in the reading table pointer 7, and the contents of the corresponding reading entry are output as the reading output 15. By sequentially selecting each reading pointer entry in the reading pointer entry selection section 6, all the readings of the kanji corresponding to the input kanji code can be obtained. It is also possible to selectively output only one reading according to instructions.

FIG. 4 shows the format of the reading pointer table 8, where 41 is the table identifier, 42 and 4
3 are fields for the first kanji code and the last kanji code of the accommodated kanji, which respectively display the upper and lower limits of the table. Each reading pointer entry column 9
The first 3 of the 6 reading pointer entries in
One can be distinguished as one for On-yomi, and the latter three for Kun-yomi.

FIG. 5 shows the format of the reading table 11, where 51 is a table identifier, and 52 and 53 are the beginning offset and the last offset of the accommodation reading, respectively, which indicate the upper and lower limits of the table.
Reading entry 12 is EBCDIC code kana character 6
It is displayed within the same number of characters, and the same reading does not exist in the table.

FIG. 6 shows a specific example of the correspondence between the reading pointer table 8 and the reading table 11. The example shown is for the kanji "ichi", and the reading table 11 shows the readings of this kanji: on-yomi "ichi", "itsu", kun-yomi "hajime", "hitotabi",
“Hitotsu” is registered, and in the corresponding reading pointer entry column of the reading pointer table 8,
Offset values for each of the above readings in the reading table 11 are set respectively. Please note that there are only two on-yomi readings, so use the empty one.
“0” is set in each reading pointer entry.

〔Effect of the invention〕

As described above, according to the present invention, the table for obtaining readings from kanji codes can be made smaller than before, so less memory space is required.
Also, most of the reading corrections can be done by changing the pointers in the table.

[Brief explanation of the drawing]

FIG. 1 is a configuration diagram of a device according to an embodiment of the present invention;
Fig. 2 is an explanatory diagram of a conventional reading table, Fig. 3 is a conceptual diagram of the method of the present invention, Fig. 4 is an explanatory diagram showing an example of the format of a reading pointer table, and Fig. 5 is an explanation showing an example of the format of the reading table. FIG. 6 is an explanatory diagram of a specific example of the correspondence between the reading pointer table and the reading table. In the figure, 1 is a Japanese information processing device, 2 is a storage device, 3 is a table access control unit, 4 is a kanji code/address conversion unit, 5 is a reading pointer table pointer, 6 is a reading pointer entry selection unit, 7
is a reading table pointer, 8 is a reading pointer table, 9 is a reading pointer entry string, 10 is a reading pointer entry, 11 is a reading table, and 12 is a reading entry.

Claims (1)

[Claims] 1. In a Japanese information processing device, a reading table means that arranges the readings of all kanji stored in the device in a certain order, and a position in the reading table means that represents all the readings of each kanji in the order of the kanji code. reading pointer table means in which one or more reading pointers indicating the reading pointer are arranged, reads out the corresponding reading pointer in the reading pointer table means based on the input kanji code, and then using the reading pointer. A kanji reading information retrieval method characterized by reading out the reading from a corresponding position in a reading table means.