JPH03176732A - Sorting processing method - Google Patents

Abstract

PURPOSE:To shorten a merge processing time by utilizing a quick sorting method in an external sorting method and setting up the length of a record string to be an intermediate sorting result longer than the number of records to be simultaneously stored in a main storage area. CONSTITUTION:Plural records capable of being stored in the main storage area are read out from the head of an original record string 1 and arranged in the prescribed strength order of keys. The records of weak keys are sorted and the record string is outputted so that an intermediate record string 5 is extended. In the case of only strong records, all the records in the area 3 are sorted and the record string is outputted to a working storage area 4. The key value of the weakest record is set up to a new reference value and unprocessed records in the string 1 are read out to the area 3 and the same processing is repeated. A merge processing part 6 executes the merge processing of the intermediate record strings and outputs a code string 7, so that the merge processing time can be shortened.

Description

[Detailed Description of the Invention] [Summary] Regarding sorting processing in a computer, in the case of external sorting, the length of the record string of intermediate sorting results is stored in the main storage area at once by effectively utilizing the quick sort method. The purpose of this is a sort processing method that efficiently makes the length of records longer than the number of records entered, and sorts the records of a ring record group in the order of a predetermined key according to the predetermined strength of the key value determined for each record. In the case of processing, the records within a predetermined number of the ring record group are read into a predetermined main storage area, and among the read records, only the records whose keys are not stronger than a predetermined reference value are arranged in the order of the keys, It is added to and concatenated with the record string that was output to the working storage area immediately before, and if there is no record with a key that is not stronger than the corresponding M quasi-value,
Arrange all records in the main storage area in the order of the key,
output as a new record string to the working storage area, set the weakest key of the record output to the working storage area as the new reference value, start with a predetermined initial value as the reference value,
After sequentially repeating the above processing for all records of the ring record group, a predetermined merge process is executed for the record string output to the working storage area.

[Industrial application field]

The present invention relates to a sorting method in which a quick sort method can be effectively used for sorting in a computer, particularly for sorting when there are a large number of records.

[Prior art and problems to be solved by the invention] As is well known, sort processing in a computer is a process of arranging the records of a ring record group in order of predetermined key strength based on the key value determined for each record. In the case of so-called internal sorting, in which records read into main memory are sorted, the quick sort method (see page 60 of "Information Processing Handbook" edited by Information Processing Society of Japan) is known as a relatively fast processing method. ing.

When the number of records becomes relatively large, it becomes impossible to read all the records into main memory at once and process them, so so-called external sorting becomes necessary. In external sorting, for example, as shown in FIG. 3, the sort processing unit 2 divides a ring record string l forming a file stored in an external storage device such as a magnetic disk, reads it into the main storage area 3, and stores it in some way. The records are partially sorted according to the method, and the record strings resulting from the intermediate sorting are sequentially output to the working storage area 4 of the external storage device.

The merge processing unit 6 then merges the intermediate record strings 5, which are generally generated in plural in this way, using an appropriate method, thereby merging them into l゛ output record strings 7 in the required key order. Steps will be taken.

In this merge process, for example, it is necessary to sequentially extract each record with a strong key from all the sorted record strings and arrange them in key order, so to speed up the merge process, make the sorted record string as long as possible. , it is desirable to reduce the number of columns.

On the other hand, in order to generate this intermediate record string, the above-mentioned quick sort method can be used to process quickly, but as mentioned above, this method can only efficiently process records that can be read into main memory at once. Since the number range is limited, the length of the record column of the sort result is limited by the size of the main storage area and is generally relatively short.Therefore, the number of intermediate record columns increases, reducing the processing amount of the merge process. There is a problem that the amount increases.

In the case of external sorting, the present invention makes it possible to effectively make the length of the record string of intermediate sorting results longer than the number of records stored in the main storage area at one time by effectively utilizing the quick sort method. The purpose is a sort processing method.

[Means to solve the problem]

FIG. 1 is a process flowchart showing the configuration of the present invention.

The figure shows the configuration of a sort processing method, in which records of a ring record group are sorted in a predetermined key order according to a predetermined strength order of key values determined for each record.
After predetermined initialization in processing step IO, processing step 11 reads the records within a predetermined number of the ring record group into a predetermined main storage area, and processing step 12.13 reads a predetermined number of records from among the read records. Only the records with keys that are not stronger than the reference value are arranged in the order of the keys, added to and concatenated with the record string that was output to the working memory area immediately before, and in processing step 12, records with keys that are not stronger than the reference value are output. If there is no record, in processing step 14 all records in the main storage area are arranged in the order of the key and output as a new record string to the working storage area, and in processing step 15, the records output to the working storage area are The weakest key is newly set as the reference value, a predetermined initial value is started as the reference value, and the above process is sequentially repeated for all records of the original record group identified in processing step 16, A predetermined merge process is executed for the record string output to the working storage area in processing step 17.

[For production]

With the above processing method, if the key of the record to be read into the main storage area has a key that is not stronger than the key of the record in the working storage area that was output immediately before, the record with such a key will be included in the read record. As long as the records continue to be stored, the records are linked into one record string in the working storage area, so the intermediate record string can be made longer regardless of the size of the main storage area.

Therefore, if the size of the original record string is N records and the main storage area is the size of G records, in the worst case where the records in the original record string are arranged in completely reverse key order, the number of columns in the intermediate record string will be will be N70 as before, but if the keys are arranged in the normal order, there will be one (that is, the intermediate record string has already been merged), and in other cases, the value will be the intermediate value between the two. Become. However, in practice, the original record string is often arranged in the normal order of the keys to some extent, so it can be expected that the value is usually much smaller than N/G records.

〔Example〕

Processing steps 10 to 16 in FIG. 1 are the flow of processing according to the present invention in the sort processing unit 2 in FIG. In this case, the area, necessary control information, etc. are set to initial values, and the reference value is set to a predetermined initial value.

Any value that the key can take may be used as the initial value of the base Ys value, but for example, the weakest possible value of the key should be used as the initial value! Let it be the tJI value.

After the above preparations, start the sorting process and process step 1.
In step 1, as many records as can fit into the main storage area 3 are read in order from the beginning of the original record string 1, and in processing step 12, they are classified into, for example, records with keys stronger than a reference value and records with keys less strong.

As a result, if there is a record with a key that is not stronger than the reference value (i.e. equal to or weaker than the reference value),
In processing step 13, only the records whose keys are not stronger than the reference values are sorted, and that record string is outputted so as to extend the intermediate record string 5 output immediately before on the working storage area.

If there are only records that are stronger than the standard value, the main storage area 3
As a new record column (separate from the previously output record column), a record column is created by sorting all the records in . Output to working storage area 4. Note that in the first process, an intermediate record string is output via this bus.

After one of the above processes, the value of the weakest key among the keys of the record output this time (the key value of the last record in the output record sequence) is set as the new reference value in processing step 15. . Then, in process step 16, identifying
If there is an unprocessed record in the original record column 1, return to processing step 11, read the next record into the main storage area 3 to replenish the number of empty records that have been output, and repeat the above processing using the new reference value. repeat.

When all the records in the original record string l are processed through the above process, as many intermediate record strings as those that have passed through processing step 14 are created in the working storage area 4. The merge process is performed on the intermediate record string of , and output record string 7 is output.

Figure 2 is a diagram explaining the above process using a simple key example. The numbers in the figure are the key values of each record, and indicate an example of sorting in ascending order of key values. The three columns from the left are the main storage areas. The working memory is generated by outputting the sorted record string, showing the state of the read key, the state of classifying it by reference value, and the state of sorting all or the lower half of the classified results. The intermediate record string on the area is shown in the rightmost column.

Figure 2 (a) is an example of processing the first 7 records.The standard value is initialized to the strongest key value, for example 99, and all the records are sorted because all the key values are stronger than the standard value. and output it as an intermediate record string.

At this time, the weakest key was "7", so the base f$ value was set to "7" and the next seven records were read as shown in the leftmost column in (b). If you classify the key values before and after the standard value "7", the result will be as shown in the second column, so the keys "10.8.9" which are not stronger than the key value "7J" will be sorted, and the keys "10.8.9" will be sorted before and after the key value "7J". Connected to the output intermediate record string.

Next, as shown in (C), only the three empty records are read into the main storage area, and the reference value is set to the weakest key "10" of the previous output record for processing.

In this example, all keys are stronger than the standard value "10", so all records are sorted and a new second key is created.
to an intermediate record column. Furthermore, if there is even one record whose key is not stronger than "10" (i.e., with a key of 10 or more) among the three records read,
As in the case of , those records are connected to the first intermediate record string, and the second intermediate record string is not yet created.

〔Effect of the invention〕

As is clear from the above description, according to the present invention, in external sort processing in a computer, the quick sort method is effectively used to determine the length of the record string of the intermediate sort result by the records that can be stored at one time in the main storage area. Since it can be efficiently made longer than the number, the time for subsequent merging processing is shortened and the overall processing efficiency is improved, which has a significant industrial effect.

[Brief explanation of drawings]

FIG. 1 is a process flowchart showing the configuration of the present invention, FIG. 2 is an explanatory diagram of a processing example, and FIG. 3 is a block diagram of an example configuration of a sorting processing system. In the figure, 1 is a ring record string, 2 is a sort processing section, 3 is a main storage area, 4 is a working storage area, 5 is an intermediate record string, 6 is a merge processing section, 7 is an output record string, 10
-17 indicate processing steps.

Claims (1)

[Claims] In the case of a process of sorting records of an original record group in a predetermined key order according to a predetermined order of strength of key values determined for each record, within a predetermined number of records of the original record group. Read the records into a predetermined main storage area (11), and among the read records, only records with keys that are not stronger than a predetermined reference value are arranged in the order of the keys and added to the record string output to the working storage area immediately before. If there is no record with a key that is not stronger than the reference value, all records in the main storage area are arranged in the order of the key and the work is output as a new record string. Output to the storage area (12, 14), set the weakest key of the record output to the working storage area as the new reference value (15), start with a predetermined initial value as the reference value, and convert the original record. After sequentially repeating the above processing for all records in the group (10, 16), a predetermined merge process is executed for the record string output to the working storage area (17).
A sort processing method characterized by the following.