JPS63241633A - Record retrieving system - Google Patents

Abstract

PURPOSE:To enhance the speed-up of a retrieval by retrieving from the position of a record retrieved just before without executing the retrieval of an index part when the record having a desirable key value is included in the records group retrieved just before in case of continuously retrieving the records whose parts of high rank of key are equal. CONSTITUTION:An index part 22 has a function for storing the position of records group including the record having the desirable key value. The records whose 'parts of the high rank of the key' are same are usually included in the same records groups R1-Rm. Therefore if the records whose 'parts of the high rank of the key' are same are continuously retrieved, it is informed that the desirable record is same as the records groups retrieved just before according to the information prescribing the positions of the above-mentioned records groups R1-Rm retrieved just before. Thus, the retrieval of the index part 22 is made needless, so that the speed-up of the retrieval can be enhanced.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention provides a method for obtaining a desired record from a set of records in which a key for identifying the record exists in the data in the record by specifying the key. This paper relates to a record search method that allows for

[Conventional technology]

In the conventional method, for example, as disclosed in Japanese Patent Application Laid-open No. 58-51348, data in a record includes a data part, which is a set of records in which a key for identifying the record exists, and the location of the record. For files consisting of an index section, which is a collection of data to be managed, the search time was shortened by devising keys to search the index section at high speed.

However, when searching continuously for records that exist in the record group searched just before, the search is performed from the beginning of the record group each time, so the search is performed to the extent that the desired record clearly does not exist. No consideration was given to the fact that this could lead to

[Problem that the invention seeks to solve]

The above-mentioned conventional technology does not have a memory function for storing the position of the record searched just before when continuously searching for data with the same "upper part of the key", so the index section is searched each time the search is performed. , the desired record is searched sequentially from the beginning of the position of the record group determined from the index section, and as a result, there is a problem in that a range where the desired record clearly does not exist is also searched.

The present invention has been made in view of the above circumstances, and its purpose is to solve the above-mentioned problems in the conventional record search method, and to solve the problems described above in the conventional record search method. If the record group searched just before includes a record with the desired key value, the search is speeded up by searching from the position of the record searched just before, without searching the index section. The objective is to provide a possible record search method.

[Means for solving problems]

The above-mentioned object of the present invention is to have a data part that is a set of records in which data in the records has a key for identifying the record, and an index part that is a set of data that manages the location of the records. In a record search system that searches a file for a record having a desired key value using data in the index section,
Providing means for storing information for specifying the position of the record group searched immediately before, and information for specifying the position of the key value searched for immediately before and the record corresponding thereto,
If it is determined based on the information stored in the storage means that a record having the desired key value exists in the record group in which the record searched immediately before is stored, the search is performed from the position of the record searched immediately before. This is achieved by a record search method characterized by starting.

[Effect]

The index section has a function of storing the position of a record group including a record having a desired key value. By the way, records with the same "upper part of the key" are usually included in the same record group.

As a result, when consecutively searching for records with the same "upper part of the key", the information specifying the position of the previously searched record group mentioned above allows the desired record to be located in the previously searched record group. It is known that it is the same as

This eliminates the need to search the index section, and furthermore, the key value searched just before and the information specifying the position of the record corresponding to this key value allow the desired record to be located before or after the previous record position. is known to exist. This eliminates the need to search from the beginning of the record group, making it possible to omit searching in a range where the desired record clearly does not exist.

〔Example〕

Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram of a record search system that is an embodiment of the present invention. In the figure, 1 is realized by a CPU, and a control unit 11. Index search section 12. Record search section 13. Storage unit (1) for specifying the position of the record group searched immediately before14. This is a record search program consisting of a storage section (2) 15 that stores the most recently searched key value and the position of the record corresponding thereto.2 is realized by a magnetic disk device, and the data section 21. A file 3 consisting of an index section 22 is realized by a memory, and has a record group position storage area 31. This is a storage area consisting of a key value and a record position storage area 32 corresponding to the key value.

The data section 2 consists of record groups R1 to Rm, which are a subset of records r1 to rn.

Further, each entry in the index section 22 exists corresponding to each record group, and consists of a key value of 100 to Kn of the last record of each record group and an address A1 to Am of the record group.

The record group position storage area 31 consists of the key value of the first record and the key value of the last record of the record group searched immediately before, and is referenced and set by the storage unit (1) 14. In addition, the key value and record position storage area 32 is
It is referenced and set by the storage unit (2) 15.

FIG. 2 is a flowchart of the record search program 1, and the operation of this embodiment will be explained below with reference to FIG.

In this embodiment, a flag indicating the start of a search is provided in the control unit 11, and this flag is initially set to be on. In the search start state, that is, in the state where the flag is on, the search operation for the record having the key value K 600 is as follows.

It is determined whether or not it is time to start a search based on the above-mentioned flag (step 31). If it is time to start a search, a record search similar to the conventional method is performed. That is, the index search unit 12 repeats the procedure (binary search method) of dividing the index section into two parts and selecting the part where the key value K 600 of the desired record exists, and then divides the index part into two parts and selects the part where the key value K 600 of the desired record exists. (step 34)
).

Next, the record search unit 13 searches the index obtained above for the record group RI indicated by the address AIO of 750.
Desired records are sequentially searched from the first record r501 of O, and the record r600 is obtained (step 35). Next, the storage unit (1) 14 sets the key value K 501 of the first record r501 of the searched record group RIO and the key value K 750 of the last record r750 in the record group position storage area 31. (Step 36). Furthermore, the storage unit (2) 15 stores 600 as the key value of the obtained record r600 and the record position AIO+Ω in the key value and record position storage area 32 (step 39).

In the above state, the search operation for records in which the "higher part of the key" has the same key value of 650 is as follows.

It is determined whether or not it is time to start a search (step 31). Since it is not the time to start a search, the storage unit (1) 14 stores 501 as the key value of the first record of the record group searched just before, and sets the key value of the last record as 501. 750 from the record group position storage area 31 (step 32).

Next, the control unit 11 sets 501. to the key value representing the position of the record group acquired above. 750 and 650 for the key value of the desired record. Here, since the desired record r650 exists in the immediately preceding search record group RIO, the comparison result is "include" (step 33).

Next, the storage unit (2) 15 obtains the previously searched key value of 600 and the record position AIO+Ω from the key value and record position storage area 32 (step 37). As a result of comparing the obtained key value of 600 with the key value of the desired record of 650, it can be determined that the desired record exists after the previous search record position, so the search is performed from the previous search record position Al0-111. Search backwards for desired records in order, and select the record r6
Get 50. If the desired record exists ahead of the previous search record position, the desired record is sequentially searched forward and the record is acquired (step 38).

Also, records r that do not exist in the previous search record group
The operation to search for loo is as follows.

First, since it is not the time to start the search (step 31), the storage unit (1) 14 stores the key value 501 of the first record of the record group searched immediately before and the key value K 750 of the last record at the record group position. The information is obtained from area 31 (step 32). Next, the control unit 11 assigns 501. to the key value representing the position of the record group acquired above. to 7
50 and the key value of the desired record is 100.

Here, unlike the previous case, the desired record rlo
Since o does not exist in the immediately preceding search record group R10, the comparison result is "does not include" (step 33),
The subsequent operation will be the same as the operation at the start of the search (step 3).
4-36.39).

FIG. 3 is a diagram showing the record structure of the data section 21, and the forward or backward sequential search method in step 38 will be explained with reference to this diagram.

When searching backward from the current record r2, the current record length ρ2 is added to the current record position adr2 to find the position adr3 of the immediately following record r3, and the operation of locating it is repeated.

When searching forward from the current record r2, the previous record length Ql is subtracted from the current record position adr2 to find the position adr1 of the immediately previous record r1, and the operation of locating it is repeated.

According to the above embodiment, when searching for a record that has the same key value and exists in the immediately previous search record group, the search in the index section can be deleted. The first record of the group (
In the above example, r 501) to the previous search record r
This has the effect that up to 600 searches can be deleted.

In addition, since it is notified that the desired record exists before or after the record searched immediately before, there is no need to search from the beginning of the record group, and searching in a range where the desired record clearly does not exist is avoided. This makes it possible to significantly shorten the average processing time for searches in processes that involve a high search frequency as described above.

In the above embodiment, in order to determine whether or not it is time to start a search, a flag indicating the start of a search is provided in the control unit 11, and this flag is set to be on as an initial setting. However, this method is different from other methods, for example, in the record group position storage area 31 of the storage area 3, the key value of the record group R1 at the beginning of the data section 21 is set to 100 in the index section 22 as a default value, and the key value is 100. A method may also be considered in which the address A1 of the record group is set.

〔Effect of the invention〕

As described above, according to the present invention, there is a data section that is a set of records in which a key for identifying the record exists in the data in the record, and an index section that is a set of data that manages the location of the record. In a record search system that uses data in the index section to search for a record having a desired key value from a file containing Means for storing a value and information for specifying the position of a record corresponding to the value is provided, and based on the information stored in the storage means, a record searched immediately before is stored in a record group. If it is determined that a record with the desired key value exists, the search is started from the record position that was searched immediately before, so records with the same "upper part of the key" are consecutively searched. When performing a search, if the record group searched just before includes a record with the desired key value, the search can be performed from the position of the record searched just before, without searching the index section. This has the remarkable effect of realizing a record search method that can speed up the search.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a record search system which is an embodiment of the present invention, FIG. 2 is a flowchart of a record search program, and FIG. 3 is a diagram showing the record structure of a data section. 1 record search program, 11: control unit, 12: index search unit, 13 record search unit, 14: storage unit (1), 15: storage unit (2), 2: file, 21:
Data section, 22: Index section, 3: Storage area, 31 record group position storage area, 32: key value and corresponding record position storage area, r1 to rn records, R1
~Rm near address, A 1~A m near address, K
100~Kn: Key value.

Claims (1)

[Claims] 1. From a file that has a data section that is a set of records in which a key for identifying the record exists in the data in the record, and an index section that is a set of data that manages the location of the record, the index section is extracted. In a record search system that searches for a record with a desired key value using the data of A record having a desired key value exists in a record group in which the record searched immediately before is stored based on the information stored in the storage means. If it is determined that this is the case, the record search method is characterized in that the search is started from the record position searched immediately before.