JPS6232495B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a final data judgment processing method, in particular, inputting a record from an external storage medium such as a magnetic disk device, and sequentially retrieving data included in the record. In a data processing device that performs processing, simply by determining the final data display indicating the final position of valid data in the input record, the final position of the record can be determined as a result, making it possible to speed up processing. The present invention relates to a final data judgment processing method.

For example, the length of a record is predetermined, and the length of valid data stored in that record is indicated by the final data display, for example, a hexadecimal code of "00". When data is retrieved sequentially from the beginning and processing is performed according to the data contents, conventionally, the end of data is determined by checking the display of the final data and using a pointer to the data currently being processed and the record final address. This was done by comparing the If all the data in the record is valid data, the final data display, for example the hexadecimal ``00'' code, is not stored, and determining the final data display alone is not sufficient to detect the end of data. be. Furthermore, as will be described later with reference to FIG. 1, even in cases where the final data display is determined indirectly using a jump table, the determination of the record final address has conventionally been omitted. I couldn't.

Figure 1 shows an explanatory diagram of the conventional method. In the figure, 1 is the record data read out on the data buffer, 2 is the final data display indicating that the subsequent data is invalid data, and 3 is the address table. 4, in which, for example, the start address (SA), end address (EA) of record data 1, and current data pointer (CP) pointing to the data currently being processed in record data 1 are stored; is a processing code indicating the data processing type; 5 is a jump table in which branch destination addresses are determined and associated according to processing code 4; 6-1 and 6-2 are data processing units for each processing code; Reference numeral 7 represents a record end processing section which performs record end processing when processing code 4 is stop code "00", that is, final data display 2.

According to the conventional system as shown in FIG. 1, after reading record data 1 from, for example, a magnetic disk device onto a data buffer, the address table 3 is initialized. Start address SA and current data pointer CP in address table 3
is made to point to the beginning of record data 1, and end address EA is made to indicate the final address of record data 1. Note that the current data pointer CP is updated every time each data is processed. Record data 1 stores data to be processed in order from the beginning, but if the effective data section is smaller than the record size, the final data display 2 indicates the end of the effective data section. This final data display 2 is set at the time of record creation. If all data in record data 1 is valid data, final data display 2 is not set. In this example, the final data display 2 is given meaning as one of the processing codes 4, and for example, hexadecimal "00" is assigned as the processing code. As shown in the processing flow shown in Figure 1B, when processing data, first the current data pointer CP and the record data 1 are
It is determined whether the end address EA and the end address EA are equal, and the end of the record is checked. This record end determination process is a process that must be performed according to the conventional system. If this judgment process is not performed, when the contents of record data 1 are all valid data and the final data display 2 does not exist, access is made to a completely unrelated memory outside the data buffer, that is, after record data 1. This is because this will lead to errors. In the record end determination process, if it is detected that the current data pointer CP and the end address EA are equal, the record end process is performed, and post-processing such as preparation for inputting the next new record or release of the data buffer is performed. When the current data pointer CP has not reached the end address EA, the data in the record data 1 indicated by the current data pointer CP is accessed, and the branch by the jump table 5 is executed according to the processing code 4 in that data. being done. Jump table 5 has each processing code 4.
For each processing code, the address of the processing destination of the data with the processing code 4 is stored in correspondence with the processing destination.In particular, when the processing code 4 is "00", it means the final data display 2, so the branch destination is stored. is the record end processing section 7. Therefore, whether or not it is the final data display 2 is determined indirectly by the jump table 5, without having to be directly determined. If it is a normal processing code, control is transferred to each data processing unit 6-1, 6-2, etc., data processing is performed there, and the current data pointer CP points to the next data of record data 1. will be updated to. Then, the process is repeated from the record end determination process described above.

Instead of using the jump table 5 in the above processing, a method is also used in which the processing codes 4 are directly judged in the order of their appearance frequency, but in any case, according to the conventional method, as described above, It was necessary to always perform record end determination processing to prevent access. Since this record end determination process must be performed for each piece of data, it affects the processing speed and poses a problem, especially in systems that have a large amount of data and need to process them at high speed.

The present invention aims to solve this problem, and aims to speed up data processing by making it possible to omit record end determination processing. Therefore, the final data determination processing method of the present invention reads a record of an arbitrary predetermined length onto a data buffer from an input/output device, and determines the first final data display indicating the end of valid data in the record. In the data processing device, the data buffer is configured to have a size that is at least the sum of the size of the record and the size of the final data display, and the data buffer is configured to have a size that is at least the sum of the size of the record and the size of the final data display. final data display setting means for setting a second final data display having the same content as the first final data display at the position; A determination means for detecting the final position of valid data in the input record or the final position of the input record by determining with a common command, and determining the end of processing for the record. It is said that This will be explained below with reference to the drawings.

FIG. 2 shows the configuration of an embodiment of the present invention. In the figure,
8 is a data processing device, 9 is an external storage device, 10 is a record, 11 is a data buffer, 12 is a final data display fixedly set on the data buffer 11, 13 is a record input section, 14 is a current data access section, 15 represents the final data display determination section.

In FIG. 2, as the data buffer 11,
Record 1 to be read from external storage device 9
At least the final data display 1 than the size of 0
An area as large as necessary for setting 2 is secured in advance. When the record 10 is read into the data buffer 11, the final data display 12 is set in advance at a location immediately after the record 11. This final data display 12
is a display having the same content as the final data display for distinguishing between a valid data section and an invalid data section within the record 10, and is expressed, for example, by a hexadecimal number "00". The record input unit 13 of the data processing device 8 issues an input/output command to the external storage device 9 in response to a data processing request,
A desired record 10 is read out to the data buffer 11. The current data access unit 14 maintains and updates a pointer to the next data to be processed in the read record 10, and accesses each data one by one from the beginning of the data buffer 11. . When one piece of data is accessed by the current data access unit 14, the final data display determining unit 15 determines whether the data content is the final data display. In this case, the determination may be made indirectly using a jump table as described with reference to FIG. When the final data display is detected, the record end processing unit 7 is activated and record end processing is performed. If the data is normal,
Each data processing unit 6-1,
6-2, . . . are activated, the data is processed, and control is returned to the current data access unit 14 for processing the next data. In this way, the above processing is repeated until the final record display is detected, but according to the method of the present invention, there is no need for the current data access section 14 to determine the final position of the record at all. Even if all the data in the record 10 is valid data and there is no final data display in the record 10, the final data display 12 preset on the data buffer 11 is detected and the record end processing section 7 This is because it is always activated and the processing of record data is completed. That is, the final data display determination section 15 only determines the final data display, and as a result, it also determines the final position of the record.

FIG. 3 shows an explanatory diagram of an application example of the present invention for an example corresponding to FIG. 1. In the figure, the symbols correspond to those in FIG. 1 and FIG. 2. In address table 3,
The start address SA of the record data 1, that is, the start address of the data buffer 11 and the current data pointer CP are held, and the end address EA is not required and is not set and held. Next to the record data 1 of the data buffer 11, a final data display 12 having the same content as the final data display 2, for example represented by "00", is set in advance. As described above, data is processed sequentially while updating the current data pointer CP.
The record end determination process shown in the process flow is not performed. If the final data display 2 does not exist in the record data 1, the final data display 12 on the data buffer 11 will always be detected, and the record end process will be started and the process will be completed. This is because the same result as directly determined using the end address EA can be obtained. Therefore, the processing speed is increased by the amount of the record end determination processing shown in FIG. 1B.

FIG. 4 is an explanatory diagram of a specific application example of the present invention, and shows the flow of processing when applied to kanji search processing when converting kana characters into kanji codes. The input information for the kanji search process is the "reading", and the output information is the kanji in the data buffer 11 corresponding to the input reading. That is, the data buffer 11 is searched to find a kanji character corresponding to the designated reading. The kanji dictionary stored in the data buffer 11 is composed of readings (1 byte x n) and kanji (2 bytes x m), and although the number of valid data in the data buffer 11 is indefinite, the final data display is 16 A stop code expressed in decimal notation "00" is used, and data after the stop code is treated as invalid data. Note that in this case, it is assumed that hexadecimal digits "00" do not exist in the valid data. In the data buffer 11, a stop code, which is the final data display, is set and held in advance at the next position of the kanji dictionary, regardless of the contents of the kanji dictionary read out from the external storage device. When a reading is specified and a request for kanji search is made, one byte is first taken out from the data buffer 11 and it is determined whether it is a stop code. If the stop code is "00", it means the end of the valid data, and therefore the process is terminated by notifying the processing requester that there is no corresponding kanji using a return code or the like. If it is not a stop code, data buffer 11
Extracts a series of yomi codes from and compares them with the specified yomi. If the readings match, the kanji code following the reading code on the data buffer 11 is set as the corresponding kanji and is reported to the processing requester. If the readings do not match, 1 byte of next data is extracted and it is determined whether the data is a stop code. If it is a stop code, it is assumed that there is no corresponding kanji character and the process is terminated, as in the case above. If it is not a stop code, it is determined whether the retrieved data is a reading code or a kanji code. For Kanji codes, the most significant bit is ``0'', and for reading codes, the most significant bit is ``1'', so this determination can be made by examining the most significant bit of the extracted data. . Kanji codes are allocated in two-byte units, so if it is a Kanji code, two bytes are skipped, a stop code is determined, and it is determined whether it is a reading code or a Kanji code. If it is a reading code, a comparison process with the specified reading is performed, and the same process is repeated.
When performing the above processing, if a stop code is not added to the end of the data buffer 11 in advance, it is necessary to determine whether the end of the data buffer 11 has been reached every time one byte is extracted from the data buffer 11. By applying the present invention, it becomes possible to detect the end of a record simply by determining the stop code, and the processing speed is increased.

As explained above, according to the present invention, it is possible to detect the end of a record simply by determining the final data display, and the processing speed can be increased by the amount of record end determination processing that is performed each time each data is accessed. can.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram of the conventional system, FIG. 2 is an implementation configuration of the present invention, FIG. 3 is an explanatory diagram of an example of application of the present invention to an example corresponding to FIG. 1, and FIG. 4 is an explanatory diagram of an example of the present invention. An explanatory diagram of a specific application example is shown. In the figure, 1 is record data, 2 is final data display, 3 is address table, 4 is processing code, 5
is a jump table, 6-1 and 6-2 are data processing units, 7 is a record end processing unit, 8 is a data processing device, 9 is an external storage device, 10 is a record,
11 is the data buffer, 12 is the final data display,
Reference numeral 13 represents a record input section, 14 a current data access section, and 15 a final data display determination section.

Claims (1)

[Claims] 1. Read a record of an arbitrary predetermined length onto a data buffer from an input/output device, and process the data while determining a first final data indication indicating the end of valid data in the record. In the data processing device, the data buffer is configured to have a size that is at least the sum of the size of the record and the size of the final data display, and at a position on the data buffer immediately after the record read onto the data buffer. 1st above
final data display setting means for setting a second final data display having the same content as the final data display; and determining the first final data display and the second final data display using a common command. and determining means for detecting the final position of valid data in the input record or the final position of the input record to determine the end of processing for the record. .