JPH03259329A - Key relative address classifying system for bulk data - Google Patents

Abstract

PURPOSE:To reduce the number of times of unification and to shorten classifying time by forming a classification record by unifying a key and a record relative address in an input file, and reducing the length of the classification record to the minimum required length. CONSTITUTION:A classification record generating part 112 forms a key part by fetching and unifying a first key and a second key from a record inputted from a record input part 111. Also, the classification record in which the record relative address in the input file when the record is inputted at the input part 111 is attached behind the key part is generated, and the classification record is developed at the classification work area 131 of a main memory device 13. A record classification part 113 classifies and re-arranges classification record groups developed in the area 131 according to the key part. A string accumulation part 114 sets re-arranged classification record groups in one string and stores it in the temporary file 123 of an auxiliary memory device 12. In such a manner, it is possible to realize the temporary file small in capacity in the device 12.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention provides a key relative address classification method for large amounts of data;
In particular, the present invention relates to a key relative address classification method for large-capacity data that requires an auxiliary storage device as a temporary device during classification processing.

[Conventional technology]

Data classification in a data processing system involves rearranging records according to certain rules using keys within the records, and is necessary for high-speed data retrieval. For this reason, it is often used in data processing systems.

For example, it is used to sort students at a certain school based on their test results and output the data.

Conventionally, in this type of classification method, for ease of processing when outputting records, classified records are classified by matching the key and record as shown in Prisoner 6, or input records as shown in Figure 2 as classified records. , expand to the classification work area in the main memory and classify.

After that, it was output as a string to a temporary file in the auxiliary storage device, the multiple strings in the temporary files in the auxiliary storage device were merged, and then only the records were output to the output file.

[Problem to be solved by the invention]

In the conventional σ) classification method described above, the classification record is matched with a key, or the input record is classified as a classification record, and is auxiliary outputted to a temporary file in the storage device. There is a drawback that the file capacity is required to be large or equivalent to the input file.

Also, since there is information in the classification record that is not necessary for classification, the number of classification records that can be expanded when this classification record is expanded to the classification work area in the main memory is reduced, so the number of records in one string is 1t + number. This method has the drawback that the number of strings decreases, the number of strings increases, and the number of times of merging increases during subsequent merging, which lengthens the classification time.

[Means to solve the problem]

The key relative address classification method for large volume data of the present invention is as follows:
A record input section that inputs records from an input file, a classification record creation section that creates a classification record that combines a key and a record relative address in the input file, and expands it to a classification work area in the main storage, and a classification work area. A record classification section that sorts the classified record group developed in the storage area according to the key, a string storage section that stores the classified record group in the classification work area as one string in a temporary file in the auxiliary storage device, and an auxiliary storage device. A string merging unit that merges multiple strings stored in a temporary file in the device, a record output unit that outputs records from an input file to an output file according to record relative addresses in a classification record, and each of the above units. A device power file consisting of a classification control unit for controlling a device, an auxiliary storage device having a temporary file for classification work, and a main memory storage device having a classification work area used for classification work. It consists of:

〔Example〕

Next, the present invention will be explained with reference to the drawings.

FIG. 1 is a block diagram of one embodiment of the present invention.

The processing device 11 includes a record input section 111, a classification record creation section 112. Record classification section 113.

String accumulation section 114l String merging section 115.

The record input unit 111 is composed of seven units (record unit 116, classification control unit 1l7o),
An input record as shown in FIG. 2 is input from the input file 121 in the .

The first key shown in FIG. The second key is a key for classification, and in this embodiment there are two keys in the record.

The classification record creation section 112 takes out the first key and the second key from the record input in the record input section 111 to form a key part, and then inputs the input when a record is input by pressing a button in the record input section 111. Create the classification record shown in Figure 3 with the record relative address in the file appended to the key part, and classify work area 1 in the main memory ttla.
The classification record is developed in 31.

The record classification unit 113 classifies the classification record group developed in the classification work area 131 according to the key part of the classification records, and rearranges them in order. String storage section 114
stores the group of classified records in the classification work area 131 that have been sorted and sorted in order as one string in the temporary file 123 in the auxiliary storage device [12].

The string merging unit 115 inputs four strings among the plurality of strings stored in the temporary file 123, merges them into one string according to the key part of the classification record, and outputs the string to the temporary file 123 again. This process is repeated until the multiple strings are turned into four strings.

The record output section merges the four strings in the file 123 into one string according to the key part of the classification record, while outputting from the input file 121 using the record relative address in the input file in the classification record. Enter the record indicated by the intra-record relative address,
The records are output to the entire output file 122.

As shown in the control flow in FIG. 4, the classification control unit 117 causes the record input unit 1111 and the classification record creation unit 112 to check whether the classification work area 131 is full or from the input file 121 by checking the classification creation area expansion end 1171. Repeat at the box where record input ends. Thereafter, control is passed to the record classification section 113 and the record storage section 114.

By checking the record input end 1172, the above operation is repeated in the case where record input from the input file 121 ends.

After that, the string merging section 115. Record output section 116
Pass control to. As described above, the record input unit 1119 and the classification record creation unit 112. Record classification unit 113 String storage unit 114 . String merging section 115. Controls each part of the record output unit 116.

The auxiliary storage device 12 stores input files 121. Output file 122. - It has a time file 123.

The auxiliary storage device is usually a disk device or the like.

The input file 121 is a file containing input records before classification. The output file 122 is a file that outputs output records after classification.

The -time file 123 is a file temporarily created for classification work, and is used to accumulate strings.

The main storage device 13 has a classification work area 131.

The classification work area 131 is an area for developing and classifying classification records.

FIG. 5 is an example of an operation in which classification according to the classification method of the present invention is applied to a small amount of data.

Since it is a small amount of data, the string storage section 11 in FIG.
4. The operation of the string merging unit 115 will be omitted.

The input file 61 is the results of a certain school's exam, which are currently arranged in alphabetical order of names. These are sorted and discarded in descending order of total points, and in the case of a tie, they are sorted in descending order of previous ranking and output to the output file 64.

The total as the first key and the previous ranking as the second key are extracted from the trench input file 61, and a classification record group 62 with an in-file relative address is created by the record input unit 111 and the classification record creation unit 112 shown in FIG. death,
Then, the record classification unit 113 creates a classification record group 63 by classifying the classification record group. Thereafter, the record output unit 116 outputs to the output file 64.

〔Effect of the invention〕

As explained above, the present invention makes it possible to shorten the length of the classification record to the minimum necessary length by making the classification record a combination of the key and the record relative address in the input file. This has the effect of making the temporary file smaller than before.

In addition, as the number of classification records that can be developed in the classification work area in the main memory increases, the number of 0 classification records in one string increases, and the number of strings can be reduced compared to the conventional method, so the number of times of merging at the merging section is reduced. decreases,
It also has the effect of shortening the classification time.

[Brief explanation of drawings]

Figure 1 (Note: A block diagram of an embodiment of the present invention, Figure 2 (Note)
Also, FIG. 3 is a diagram showing an input record in an embodiment of the present invention, FIG. 3 is a diagram showing a p-kero classification record in an embodiment of the present invention, and FIG. 4 is a diagram showing a side division flow of the classification control unit. , FIG. 5 is a schematic diagram showing an example of the operation of the present invention, and FIG. 6 is a diagram showing a classification record in a conventional example. 11...Processing device, 12...Auxiliary storage device, 13...Main storage device, 111...
-Record input section, 112... Classification record creation section, 113... Record classification section, 114...
. . . String storage section, 115 . . . String merging section, 116. ...Record output section, 1
17...Classification control unit, 121...Input file, 122...Output file, 123.
・・・・・・-Time file, 131...Classification work area, 1171...Classification work area expansion complete, 1
172...Record input completed, 61...
・Input file example, 62... Classification record example,
63...Example of classification record for minute at, 64...
...Output file example.

Claims (1)

[Claims] a record input section that inputs records from an input file; a classification record creation section that creates a classification record that combines a key and a record relative address in the input file and deploys it in a classification work area in a main storage; and the classification work area. a record classification unit that classifies classified record groups developed in the storage area according to keys; a string storage unit that stores the classified record group in the classification work area as one string in a temporary file in an auxiliary storage device; a string merging unit that merges a plurality of strings stored in a temporary file in an auxiliary storage device; a record output unit that outputs a record from an input file to an output file according to a record relative address in the classification record; and each of the above units. A large storage device that has a classification control unit that controls the data, an auxiliary storage device that has input files, output files, and temporary files that are temporary files for classification work, and a main storage device that has a classification work area that is used for classification work. Key-relative address classification scheme for capacity data.