JP2001117929A - Data retrieving method, data aligning method and data retrieving device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform completely coincident retrieval without dispersion in retrieval time and high-speed ambiguously coincident retrieval to a data structure, which defines key data in records as retrieval data. SOLUTION: This device is provided with a retrieval object data storage part 4 for storing data to become a retrieval object for each digit in a bisected tree structure on a memory and a data retrieving part 5 for retrieving the target data from the route of the stored data in the bisected tree structure by comparing the data of respective nodes with the retrieval data.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a high-speed search and sorting of character strings while storing data to be searched in a memory in a binary tree structure for each digit and comparing the search data digit by digit. The present invention relates to a data search method, a data alignment method, and a data search device.

[0002]

2. Description of the Related Art As is well known, in a binary tree search, data to be searched is expressed in a binary tree structure, and search data is sequentially compared from root data to search for corresponding data. Is what you do.

Here, the tree structure is a data structure for expressing a hierarchical relationship of data. In the hierarchical relationship, when an element B exists below a certain element (node) A, a parent and a child are expressed by giving a vertical relationship between the elements. If the connection between parent and child is defined as a branch (path), an element without a parent is defined as a root (root), an element without a child is defined as a leaf, and an element without a parent or child is defined as a node, as shown in FIG. Such a hierarchical relationship can be created. Further, the binary tree structure refers to a tree structure having two child elements for a certain parent.

[0004] With respect to this tree structure, with a certain element as a root, data smaller than the root is stored on the left and large data is stored on the right. The magnitude relation of each data is represented by data "a, b, c,
d, e, f, g], a <b <c <d <e <
From f <g, the tree structure is as shown in FIG.

In general, the number of data comparisons at the time of retrieval in this binary tree is as follows, where n is the number of data. The average number of comparisons: log ₂ n The maximum number of comparisons: n There are two types of searches: perfect match search and fuzzy search (also called partial match search). In an exact match search, for example, 1
If the data is 0 bytes, the search data and the search target data are searched by comparing all 10 bytes. In the fuzzy search, if the search data is 20 bytes, a hit is made if the 10-byte data matches.

[0006]

However, in a perfect match search, the maximum number of data comparisons becomes n, which is the same as the number n of data, and the search time varies depending on the number of data. There was a problem.

[0007] In addition, in the fuzzy search, if it is attempted to search for data such as a character string which has a plurality of byte sizes, it takes a long time to search using a simple binary tree structure.

The present invention has been made in view of the above circumstances, and realizes a perfect match search and a high-speed fuzzy match search with no variation in search time for a data structure in which key data in a record is used as search data. It is an object of the present invention to provide a data search method and device capable of performing the data search.

It is another object of the present invention to provide a data alignment method capable of executing data alignment processing easily and in a short time.

[0010]

In order to achieve the above object, according to the first aspect of the present invention, data to be searched is stored in a memory in a binary tree structure for each digit, and the stored binary tree is stored. It is characterized in that search data is compared digit by digit with respect to structured data to search for target data.

According to a second aspect of the present invention, in the data search method according to the first aspect, when performing an ambiguous search, the search data is compared with the stored binary tree structure data digit by digit. Then, the process of retrieving the target data and then retrieving the data by shifting the digit by one digit is repeated until the last digit is executed.

According to a third aspect of the present invention, data to be sorted is stored in a memory in a binary tree structure for each digit, and the stored binary data is stored.
A backward search or a forward search is performed in order from the upper digit or lower digit of the data of the branch tree structure, and the data to be sorted is sorted in descending order or ascending order.

According to a fourth aspect of the present invention, there is provided a search target data storage means for storing data to be searched in a memory for each digit in a binary tree structure, and each node from a root of the stored binary tree structure data. And data search means for searching for target data by comparing the data with the search data.

According to a fifth aspect of the present invention, in the data search apparatus according to the fourth aspect, the data structure of each node constituting the binary tree includes digit data for storing the data to be searched for each digit. A pointer to digit data smaller than the own data stored in the digit data portion, a pointer to digit data larger than the own data, a pointer to the next digit data of the own data, and an end of data. And a flag indicating the flag.

According to a sixth aspect of the present invention, in the data search device according to the fourth or fifth aspect, the data search means comprises:
When performing an ambiguous search, the search data is compared digit by digit with respect to the stored binary tree structure data to search for the target data, and the process of searching again by shifting the digit by one digit is performed last. It is characterized by including a function of repeatedly executing up to the digit.

According to the above arrangement, the data to be searched is stored in the memory for each digit in a binary tree structure, and the search data is compared with the stored binary tree structure data for each digit. By searching for the target data by using the key data in the record as the search data, it is possible to realize a complete match search and a high-speed ambiguous match search with no variation in search time for a data structure in which key data in a record is used as search data.

The data to be sorted is stored in the memory in a binary tree structure for each digit, and a backward search or a forward search is executed in order from the upper digit or lower digit of the stored data of the binary tree structure. By arranging the data to be sorted in descending order or ascending order, the data sorting process can be executed easily and in a short time.

[0018]

FIG. 1 is a block diagram showing a configuration of an embodiment of a data search apparatus according to the present invention.

The data search device 1 shown in FIG.
CP consisting of a personal computer and executing a data search
U2 and an external storage medium such as a hard disk,
A data storage unit 3 for storing data to be searched.

The CPU 2 stores a search target data storage unit 4 for storing data to be searched in a memory for each digit in a binary tree structure, and stores a binary tree structure data stored in the search target data storage unit. A data search unit 3 that compares the data of each node with the search data from the root and searches for the target data
It has.

The retrieval target data storage unit 2, as shown in FIG. 2, the search target data _{is, A n A n-1 ···} A 2 A 1 · · · · X n X n-1 ··· X _{In the case of 2} X ₁ Y _n Y _n-1 ... Y ₂ Y ₁ , as shown below, data for each digit is extracted and stored in the memory in a binary tree structure for each digit. .

[0022] 1 digit _{......... A 1 ··· X 1 Y 1} 2 digit _{......... A 2 ··· X 2 Y 2} · · n-1 digit ... A _{n over} 1 ··· _{X n over} 1 Y _{n over 1} n-th digit ......... _{a n} ··· _X n _Y n in this case, as shown in C language structures in Figure 3, the search target data storage unit 2 of the memory, each For each digit, a pointer indicating the address of a buffer storing digit data, digit data smaller than the digit data, a pointer indicating a buffer address storing digit data larger than the digit data, and A pointer indicating the address of the buffer in which the next digit data of the data is stored, and "ON" if the digit data is end data (the end of the digit and no digit to be linked next), not the end In this case, there are provided areas for storing flags that are “OF”.

Next, the operation of this embodiment will be described.

As shown in FIG. 4, "142", "a42", "942", "13", "8"
It is assumed that there are nine data of “3”, “b3”, “33”, “2”, and “5.” The search target data storage unit 4 stores these nine data in a binary tree for each digit. As shown in Fig. 4, the buffer address 1 of the first digit is "3" as digit data and "address 2" is a pointer to small digit data, as shown in Fig. 4. In addition, “address 3” is stored as a pointer to the large digit data, “address 4” is a pointer to the next digit data, and “OF (abbreviation of OFF)” is stored as a flag indicating the end of the data. Has a different structure depending on the data storage order, but the example of FIG. 4 shows a case where data is stored starting from “3”.

Similarly, address 2 has "2" as digit data, "NL" as a pointer to smaller digit data,
“NL (abbreviation of NULL pointer)” is stored as a pointer for large digit data, “address 4” is stored as a pointer to the next digit data, and “ON” is stored as a termination flag. Address 3 has "5" as digit data, "NL" as a pointer to small digit data,
“NL” as a pointer for the large digit data, “NL” as a pointer to the next digit data, and “O” as a termination flag
N "are stored.

Next, at address 3 of the second digit buffer, "8" is used as digit data, "address 5" is used as a pointer to small digit data, "address 6" is used as a pointer to large digit data, and the next digit is used. “NL” is stored as a pointer to data, and “ON” is stored as a termination flag. The address 5 of the second digit has “3” as digit data, “address 7” as a pointer to small digit data, “NL” as a pointer to large digit data,
“NL” is stored as a pointer to the next digit data, and “ON” is stored as a termination flag. In addition, 2
The address 6 of the digit has “b” as digit data, “NL” as a pointer to small digit data, “NL” as a pointer to large digit data, “NL” as a pointer to the next digit data, and a termination flag. Are stored as “ON”. Further, the address 7 has "1" as digit data and "N" as a pointer to small digit data.
"L", "NL" as a pointer to the large digit data, "NL" as a pointer to the next digit data, and "ON" as a termination flag, respectively. , "NL" as a pointer to small digit data, "NL" as a pointer to large digit data, "Address 10" as a pointer to the next digit data, and "OF" as an end flag.

Next, the buffer address 10 of the third digit has "9" as digit data, "address 11" as a pointer to small digit data, "address 12" as a pointer to large digit data, and the next digit data. "NL" is stored as a pointer to "", and "ON" is stored as a termination flag. The address 11 has "1" as digit data and "N" as a pointer to small digit data.
"L", "NL" as a pointer to the next digit data, "NL" as a pointer to the next digit data, and "ON" as a termination flag, respectively. , "NL" as a pointer to small digit data, "NL" as a pointer to large digit data, and "N" as a pointer to the next digit data.
L, and “ON” are stored as the termination flag.

When the search target data is stored in a binary tree structure for each digit in this way, the data search unit 5 compares the search data and the search target data for each digit and executes the search in order from the root. Search and output the desired data. Hereinafter, the search performed by the data search unit 5 will be described by dividing it into a perfect match search and an ambiguous search (partial match search).

(1) Exact Match Search <Exact Match Search Method in the Present Embodiment> In the exact match search method, the search is executed sequentially from the root of the binary tree structure of the lower digit. If the character is found, it starts searching for the next column of characters linked to that character. Since then
This process is repeated until the most significant digit of the search target data is executed. If no data is hit, the process ends.

FIG. 5 shows how data is stored in a binary tree structure in order. When data is stored in this manner, for example, “ahd” is used as search data.
When searching from the data structure of FIG. 5, first, the search is started by comparing the first digit "d" of the key data with the root "c" of the first digit of the search target data. Since the data “d” is larger than the root data “c”, the process proceeds to the right branch and is compared with the data “e”. The key data “d” is
Since it is smaller than the data "e", it is compared with the left data "d", and here, the data of the first digit is hit.

Next, the search for the key data "h" of the second digit is started and compared with the root data "g". As a result, the data proceeds to the left branch and the data "h" is hit. Become.

In the search for the third digit key data "a", the data "b" linked to the second digit data "h" is retrieved.
As a result, the data "a" is hit by proceeding to the left branch. In this way, the search data "ahd" is searched at high speed from the binary tree structure search target data stored for each digit.

<Effects of Perfect Match Search in the Present Embodiment> In perfect match search, the number of comparisons is not proportional to the number of data but to the type of data of each digit.

In general, X_nX_n-1... X₂X₁Day of
Is M_nM_n-1... M ₂M₁Kind of day
Data, the number of data is M_n× M_n-1×
... × M₂× M₁And the number of comparisons is
is there.

[0035]

[Table 1] Since the logarithm has the property of log _a XY = log _a X + log _a Y, the average number of comparisons in the case of the conventional binary tree and the case of the present embodiment is the same.

When a large amount of data exists, the value of M is considered to exist uniformly, so that _Mn = _Mn-1
=== M ₂ = M ₁ = M

[0037]

[Table 2] Since the maximum number of comparisons in this embodiment can be smaller, a stable search can be performed without variation in search time.

(2) Fuzzy Search <Fuzzy Search Method in the Present Embodiment> Next, a fuzzy search for searching whether or not a k-digit search character (string) exists in a part of n-digit data. explain. In fuzzy search,
A search is performed for all data to be searched, and all search data is extracted. Now, the search target _{data: X n X n-1 ···} X k + 1 X k ··· X 3 X 2 X 1 search data: as _{Y k + 1 Y k ··· Y} 3 Y 2 Y 1, each digit When the search data is stored in the memory in a binary tree structure, a tree structure as shown in FIG. 6 is obtained. The processing for storing this data is the same as in the case of the perfect match search. In FIG. 6, a portion surrounded by a square indicates a united tree structure of each digit.

In this fuzzy search, for example, search data 1
If it is 0 bytes and the search target data is 20 bytes, if 10 bytes are hit, all are searched. Therefore, even if a 10-byte hit occurs first, there is a possibility that there is a hit in the next 10 bytes. For this reason, when the search data has k digits, as shown by the dashed line in FIG.
The data is searched in a range shifted by a digit and surrounded by a broken line in the figure. In this manner, the search is executed by shifting the digits sequentially to the last digit of the search target data.

More specifically, the data of n digits is now
Each X_nX_n-1... X₂X ₁, And M_n
M_n-1... M₂M₁Field composed of different types of data
Think about a match. However, X is an arbitrary value.

In order to realize a fuzzy search using a binary tree,
Since all data must be searched, the number of searches is the same as in the sequential search.

In the fuzzy search of the present invention, a search is performed according to a search range and a search order as shown in FIG.

[0043] In other words, the search target _{data: X n} X _n-1
.., X _{k + 1} X _k X _k−1 ... X ₂ X _{1 In} the first search, 1 to k digits (X _k.
2 X ₁ ) is searched, and in the second search, the data is shifted by one digit from 2 digits to k + 1 digits (X _{k + 1} X _k X
The search up to _k-1 ... X ₂ ) is executed. In this manner, the search up to the (n−k + 1) th search is executed.

<Effect of Fuzzy Search in the Present Embodiment>
In the case of the fuzzy search described above, the number of comparisons in each search range is as follows.

That is, in the first search, "KM"
Times, the second search is “M · KM (= KM ² ),
・ In the (n−K + 1) -th search, “M ^n−K KM
^{(= KM n -K +} 1) becomes ".

When a large amount of data exists, the value of M is considered to exist uniformly, so that _Mn = _Mn-1 =
.. = M ₂ = M ₁ = M, and the number of comparisons is as follows when the search data has k digits.

[0047]

[Table 3] As can be understood from Table 3 above, the fuzzy search in the present embodiment can perform a search faster than the conventional binary tree search.

(3) Data Sorting Process (Sort) A method of storing data to be searched for each digit in a memory in a binary tree structure and searching the stored binary tree structure data from the upper digit. , Data alignment processing can be realized.

Now, the data stored in the tree structure is These data are stored in a tree structure as shown in FIG. FIG. 8 shows a case where data is sorted in ascending order and a case where data is sorted in descending order.

First, in order to arrange data in ascending order,
A backward search is executed according to the scanning path indicated by 1 in the figure. As a result, since the storage position of the data can be known, the data is acquired sequentially from the first digit, as indicated by the black circles in the figure.

On the other hand, when searching for data in descending order,
A forward search is executed according to the scanning path indicated by 2 in the figure. As a result, since the storage position of the data can be known, the data may be obtained in order from the upper digit, as indicated by a black circle in the figure.

As described above, according to the present embodiment, the data alignment process can be executed easily and in a short time.

[0053]

As described above, according to the data search method and the data search apparatus of the present invention, perfect match without variation in search time is obtained for a data structure in which key data in a record is used as search data. A search and a high-speed fuzzy match search can be realized.

According to the data alignment method of the present invention,
The data sorting process can be executed easily and in a short time.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of a data search device according to the present invention.

FIG. 2 is an explanatory diagram showing a state in which a large number of search target data are grouped for each digit.

FIG. 3 is an explanatory diagram showing each element stored in a binary tree structure for each digit in a C language structure.

FIG. 4 is an explanatory diagram showing a data structure stored in a binary tree for each digit.

FIG. 5 is an explanatory diagram showing an example of a binary tree structure stored for each digit.

FIG. 6 is an explanatory diagram of a case where an ambiguous search is executed in a binary tree structure stored for each digit.

FIG. 7 is an explanatory diagram showing a search range and a search order in an ambiguous search.

FIG. 8 is an explanatory diagram showing a data alignment method according to the present invention.

FIG. 9 is an explanatory diagram of a tree structure.

FIG. 10 is an explanatory diagram showing a data search in a tree structure.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Data search device 2 CPU 3 Data storage part 4 Search target data storage part 5 Data search part

Claims (6)

[Claims] 1. A data to be searched is stored in a memory in a binary tree structure for each digit, and search data is compared digit by digit with respect to the stored binary tree structure data to obtain target data. A data search method characterized by searching. 2. The data search method according to claim 1, wherein when performing an ambiguous search, the search data is compared digit by digit with respect to the stored binary tree structure data, and the target data is determined. A data search method comprising: performing a search, then shifting the digit by one digit, and performing a search again to the last digit. 3. The data to be sorted is stored in a memory in a binary tree structure for each digit, and a backward search or a forward search is executed in order from the upper digit or the lower digit of the stored data of the binary tree structure. And sorting the data to be sorted in descending order or ascending order. 4. Search target data storage means for storing data to be searched in a memory in a binary tree structure for each digit, data of each node from a root of the stored data of the binary tree structure, and search data. And a data searching means for searching for target data by comparing the data with the data. 5. The data search device according to claim 4, wherein the data structure of each node forming the binary tree includes digit data for storing the data to be searched for each digit, and a digit data portion. A pointer to digit data smaller than the own data stored in the pointer, a pointer to digit data larger than the own data, a pointer to the next digit data of the own data, a flag indicating the end of the data, A data search device comprising: 6. The data search device according to claim 4, wherein the data search unit executes an ambiguous search,
The search data is compared digit by digit with respect to the stored binary tree structure data to search for the target data, and the digit is set to 1
A data search device, including a function of repeatedly executing a process of performing a search by shifting digits to the last digit.