JPS6393033A - Data retrieving system - Google Patents

Abstract

PURPOSE:To retrieve data at a high speed by producing a sub-keyword having the shorter bit length than each keyword to write it to each corresponding data block and having an access to the keyword after converting it into a sub-keyword. CONSTITUTION:A data table 11 of a data retrieving device 10 contains data blocks 11-1-11-n. A block 11-i includes an area KEY-i of a normal keyword (KEY) and an area DATA-i of the data (DATA) corresponding to the area KEY-i together with an area key-i of a sub-keyword (key). The keyword (key) is produced previously via a converting part 12 for each KEY and added to the head of the block 11-i. When a data retrieving request is received, the KEY for access is converted into the key for access by the part 12 for access of each block 11-i. In such a way, data can be retrieved at a high speed.

Description

[Detailed description of the invention] [Summary] When searching for data corresponding to each keyword, sub-keywords for each keyword are created in advance according to certain rules and added to each keyword. Each time a request to search for corresponding data by specifying a keyword occurs, a corresponding sub-keyword is calculated based on the above-mentioned certain rules, and the target data is searched using the calculated sub-keyword. This aims to shorten data search time.

[Industrial application field]

The present invention relates to a data search method.

The database must be accessed by specifying the desired keyword. For example, if the database is a phone number, enter the desired person's name (keyword) and enter the phone number (keyword).
data). Alternatively, if the database is, for example, a library catalog, the desired book name or author name (keyword) is input to obtain the library catalog (data). The present invention relates to a data search method using such keywords.

[Conventional technology]

In conventional data searches, keywords are compared character by character to detect whether the keyword matches a desired keyword. For example, in the example above, the person's name (temporarily “E
When reading out the telephone number (data) corresponding to EM I KA'' from the data table, first
For example, convert KA character by character using ASCII code and get "45".

4D", "49", "4B", and "41".

Then, among the keywords attached to each data in the data table, keywords that match these "45" . . . "41" are detected. In this case, each of the five characters is compared. If a match is found, the data (telephone number) is searched.

[Problem that the invention seeks to solve]

In the conventional data search described above, a keyword of N characters (N is a natural number) is input to detect the same keyword in a data table. The number of characters of the keyword in the data table varies for each keyword, but in any case, at least N character comparison operations are required for the N characters of the input keyword. therefore,
There is a problem in that it takes a considerable amount of time to access the desired keyword, and as a result, it takes a long time to search for the desired data.

[Means for solving problems]

FIG. 1 is a diagram schematically showing the principle configuration based on the system of the present invention. In the figure, numeral 11 in the data search device 10 is a data table, which constitutes a database. The data table consists of n data blocks 11-1.
11-2...11-1...11-n. Each data block has the same configuration; for example, data block 11-1 is shown in detail. The data block 11-1 is
The sub-keyword (key) area key-i, which is a feature of the present invention, comprises at least an area KEY-i for a normal keyword (KEY) and an area DATA-1 for data (DATA) corresponding thereto. i is further added.

This sub keyword key is for each keyword KEY,
Each data block (
11-i).

On the other hand, when a data search request is made, a desired keyword KEY for access is input from the lower right of the figure. This is converted into the subkeyword ke for access via the conversion unit 12.
y, and each data block (11-i) in the data table 11 is accessed using this key.

[For production]

The conversion unit 12 for generating a sub-keyword key from the keyword KEY converts the keyword KE based on certain rules.
Subkeyword ke with a bit length shorter than the bit length of Y
y, and the access keyword KE
Y and the keyword area (
KEY-i)
The time required to detect a match or mismatch between the access subkeyword key and the subkeyword area (key-i) of each data block (11-i) is shorter.

In other words, data search time is reduced. Note that the upper converter 12 and the lower converter 12 in FIG. 1 may be the same, or may be separate as long as the above-described certain rules are the same for both.

Compared to the bit length of the keyword KEY, the subkeyword ke
Since the bit length of y is short, the number (i) of data blocks (11-i) that can be distinguished by the sub-keyword key naturally decreases. Therefore, there may be cases where two or more data blocks are accessed by one access subkeyword key. At this time, the keyword K
Look at the matches and discrepancies between the EYs and identify the desired one.

〔Example〕

FIGS. 2A and 2B are flowcharts showing an example of the operation in the conversion unit 12.
This is a specific example. According to a certain rule according to this example, (a) the bit string representing the first character constituting the keyword (KEY) is shifted by 1 bit in a certain direction, and (U) the bit string shifted by 1 bit and the bit string that is shifted by 1 bit cause overflow. The operation procedure of adding the 1-bit carry and the bit string representing the second character is taken as one unit, and this is repeated for each character until the final character.

According to the above example (EMI KA), the operations shown in FIG. 3 are performed. FIG. 3 is a diagram illustrating certain rules used in the present invention with practical examples.

Moreover, FIG. 4 is a diagram showing a simplified example of FIG. 3,
Furthermore, the sub-keyword key (BA in this figure), which is the final result for the keyword (EMIKA), is also shown. In Figure 3, the first letter (first letter E
)Wl, the addition bit string (A
Since there is no DO) or carry (CRO), there is no ADO or CR.
O is also preset to O. Therefore, the addition bit string ADI for the first character W1 is the first character W1 itself.

Next, ADI is shifted by 1 bit (left shift) to obtain a carry CRI (O in this case), and a bit string ADI' shifted by 1 bit is obtained.

These CRI and ADI' are the second character W2 (EMIK
(indicating M of A) to obtain an addition bit string AD2.

AD2 is further shifted by 1 bit to obtain AD2' and carry CR2 (1 in this case) which has overflowed at this time is obtained.

AD2' and CR2 are added to the third character (■ of EMIKA) W3, and the same operation is repeated sequentially until the final character (A).

The above operating procedure is illustrated in flowchart form in Figures 2A and 2B.

To show it more clearly (using ASCII code), the fourth
As shown in the figure, the final result BA is the subkeyword ke for the keyword KEY (EMIKA).
It is obtained as y. In addition, in FIG. 4, different from the case of FIG. 3, each value (ADO, CRO.

Wl...) is shown in a flow from left to right and from the top to the bottom.

The above-mentioned fixed rule is just an example, but simulations have confirmed that the matching rate is quite low.

The match rate is the number of keywords that are accessed repeatedly by one sub-keyword, and it goes without saying that the smaller the number, the better.

〔Effect of the invention〕

As explained above, according to the present invention, conventionally, for example, a 5-byte keyword was checked five times for each byte (for each character), but it is converted into a 1-byte sub-keyword according to a certain rule. We aim to only require one test. Therefore, it becomes possible to access the corresponding data block at high speed.

In this case, two or more keywords KEY may be accessed redundantly by one sub-keyword key, but only in that case it is sufficient to perform matching between the original keywords.

[Brief explanation of the drawing]

FIG. 1 is a diagram schematically showing the principle configuration based on the method of the present invention, FIGS. 2A and 2B are flowcharts showing an example of an operation in the converter 12, and FIG. 3 is an actual diagram showing a certain rule used in the present invention. FIG. 4 is a diagram illustrating the example of FIG. 3 as a cylindrical element. DESCRIPTION OF SYMBOLS 10... Data search device, 11... Data table, 11 1.11-2~11-i~1l-n... Data block vine 12... Conversion unit, KEY... Keyword, key. ...Sub keyword, DATA...Data.

Claims (1)

[Claims] 1. A data table (11-i) comprising a plurality of data blocks (11-i) each including a keyword area (KEY-i) and a data area (DATA-i) corresponding to the keyword area (KEY-i). ) by using the access keyword (KEY), the corresponding data (DATA
), the sub-keyword (k
ey), and each keyword (
A subkeyword (KEY) that has a bit length shorter than
key) is generated in advance and each corresponding data block (1
1-i) When the sub keyword area (key-i) is written and the access keyword (KEY) is input,
This is converted into a corresponding sub-keyword (key) based on the above-mentioned certain rules, and each data block (11-i) is
and access the area (k
A data search method characterized by reading desired data (DATA) from a data area (DATA-i) corresponding to ey-i). 2. When there are two or more data blocks (11-i) having the same sub-keyword (key), further access the keyword area (KEY-i) in the data block and use the access keyword (KEY) Desired data (D
ATA). 3. The above-mentioned certain rule: (a) Shifts the bit string representing the first character of the keyword (KEY) by 1 bit in a certain direction, and (b) Shifts the bit string obtained by the 1-bit shift and that 1 bit. Add the 1-bit carry that overflowed due to the shift and the bit string representing the second character of the keyword (KEY) to obtain an addition bit string, and (c) perform the same operations as in (a) and (b) above for the added bit string. 2. The method according to claim 1, which comprises an operating procedure of sequentially repeating the steps from the third character to the final character.