JPH0528194A - Data access system - Google Patents

Abstract

PURPOSE:To reduce the data size and to maintain high speed at the same time by embedding compressed key information in an index with chain. CONSTITUTION:Data structure of an English/Japanese electronic dictionary system is composed of a hash table 11, an index with chain 12, and a real data file 13. A chain pointer is a pointer the corresponding key to the next index record, and a key identifier exists just after the chain pointer, used for check whether or not an input key corresponds to a registration key. Thus, it is not necessary to refer to the real data on the key other than an entry word for checking a collision by introducing the key identifier in the index 12 with chain. Thus the size of the data is reduced and the retrieval can be kept with high speed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data access method, and more particularly to a collision processing method of a chained hash method.

[0002]

2. Description of the Related Art Conventionally, the hash method is well known as a high-speed data structure in key-based data retrieval (reference: A. Aho et al., Translated by Yoshio Ohno, "Data Structure and Algorithm" [ Information Processing Series 11],
Baifukan (1987)). In the hash method, handling at the time of key collision becomes a problem, and various methods have been proposed as solutions. The hash method is roughly classified into an open hash method and a hash method with a chain. Generally, when there are many collisions, the hash method with a chain is adopted. In particular, when dealing with dictionary-like data, there are many collisions by nature, and the keys are often not unique. In addition, such dictionary-like data is often very large in scale, and due to the limitation of the main memory capacity, etc., since the substance of the data is stored in the secondary memory such as a file, the chained hash method is used. There is.

[0003]

By the way, in the hash method with chain, in the case of collision, the chain is traced and checks are sequentially performed by comparing with the desired key. How to hold the key information for this comparison Is a problem. For example, if the key information is included in the chain, high speed performance is maintained, but since the data size is originally large, the index information becomes huge. Also, if the chain has only pointers and the data entity is referenced, access will be slow. In particular, this problem becomes noticeable when an index for reverse lookup is held, and insertion and deletion become slower. The above-mentioned reference "Data Structure and Algorithms" also describes the problem of using such a secondary index.

Generally, an electronic dictionary typified by an English-Japanese dictionary has headwords of tens to hundreds of thousands of words, and if indexes including derived words and meanings other than headwords are also included, the data size becomes very large. growing. In addition, since it is often used interactively, high speed search is required. Therefore, in the current data access method, there is no choice but to compromise in the form of selecting either speed or capacity according to the use and the limitation of hardware.

An object of the present invention is to provide a data access method capable of simultaneously realizing reduction of data size and maintenance of high-speed search when checking a key collision in a hash method for dictionary data. And

[0006]

In the data access method according to the present invention, a plurality of search keys, a hash table,
A data structure having a chained index and an actual data file is provided, and a key identifier indicating information about a key is added to a chain pointer in the chained index.

The key identifier has only information necessary for identification with respect to a normal search key. If the key identifier is the same as the immediately preceding key in the chain link, it may be omitted. Further, when the target data has a name and the search key is the name itself, the key identifier can be designated to refer to the name of the actual data. Similarly, if the target data has a name and the search key can be interpreted from the name, the key identifier can be specified to interpret and refer to the name of the actual data.

[0008]

In the above data access method, first, the hash value of the given search key is obtained, and the contents of the corresponding position in the hash table are read. Here, when a key identifier is added to the chain pointer, it is determined whether or not the current index record corresponds to the key based on this key identifier. When the index record corresponds to the key, the data pointer is read, the data record is fetched from the position in the actual data file, and it is added to the result list.

In this way, by embedding the information about the key in the compressed index as the key identifier in the chained index, the key information for comparison at the time of a key collision can be efficiently used, and the data size can be reduced. It is possible to realize high speed search and search at the same time.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment in which the data access system according to the present invention is applied to an English-Japanese electronic dictionary system will be described below.

FIG. 2 is a block diagram showing a schematic configuration of the English-Japanese electronic dictionary system. This computer system includes a CRT 21 that displays various data and the like on a display screen, and a CRT driver 22 that controls the display on the CRT 21.
And a keyboard 23 for inputting commands, character strings, and numerical values, and a mouse 24 as a pointing device.
A keyboard / mouse driver 25 that outputs various data by a user operating the keyboard 23 and the mouse 24; a disk device 26, a disk device driver 27, a main storage device 28, a CPU (central processing unit) 2
9 and 9.

The disk device 26 is a secondary storage device for storing a large amount of data, and stores an index with a chain, an actual data file, etc., which will be described later. Data input / output of the disk device 26 is controlled by the disk device driver 27.

The main memory 28 stores an application program and data such as characters and numerical values input from the keyboard 23 and the mouse 24, as well as a hash table described later.

The CPU 29 is a circuit that controls the entire system and performs arithmetic processing on predetermined data based on various commands, and executes data retrieval processing based on a flowchart described later.

An outline of the data structure in the English-Japanese electronic dictionary system is shown in FIG. The data structure of FIG. 1 is basically a hash table 11, a chained index 12,
It is composed of three actual data files 13. As shown in the figure, in addition to the search from normal entry words (a, A, etc.), the meaning of a word (one,
It is possible to search by using the (a sound, etc.) or pronunciation katakana notation (d, a, etc.), so that the number of keys / records is very large, and the record for a certain key is not unique.

First, the hash table 11 holds a pointer to the chained index 12 in 3 bytes from the address indicated by the hash value h (k) of the search key k. If the corresponding key is not registered, FFFFFFH is held.

Next, details of the chained index 12 will be described. The chained index 12 is a set of index records having information corresponding to the data records on a one-to-one basis. The structure of the index record is shown in FIG. The index record includes chain pointers 1 to n related to all the keys set in the record,
It holds an array of pairs of key identifiers 1 to n and a pointer (data pointer) to a data file.

The chain pointer is a pointer to the next index record related to the corresponding key, and a list (registered data) of registered keys having the same hash value is formed by collision. The head of each list is directly pointed to by the hash table. When there is a collision, each pointer holds the address of the next chain pointer as shown in FIG. 4, and at the end of the list, nil is 0000.
00H is stored. The chain pointer is represented by 3 bytes and can take a value of 000000H to 7FFFFFH.

The key identifier exists immediately after the chain pointer and is used to check whether the input key corresponds to the registration key. The description rule of the key identifier in this embodiment is shown below. The character code is EUC.

(1) In a normal search key, the lower 1 byte of each character code of the character code string of the registration key and 00
The byte string of the logical sum of 80H is stored in order.

(2) If the registration key can be derived from the entry word, 80H is stored. This applies not only when the key is the entry word itself, but also when "A, a" or "colo
Headwords such as (u) r ”can be interpreted by making general notations into rules.

(3) It is omitted when the registration key is the same as that immediately before the chain.

In this way, by introducing the key identifier, it is not necessary to refer to the actual data by the key other than the entry word for checking the collision. However, in that case, the information amount of the registration key is low. This makes it impossible to distinguish when the lower 1 byte of the key is the same and the hash values are the same, even if the keys are different, but it is considered that there is no practical problem if the hash function is carefully examined. Also, in an electronic dictionary that only searches,
Since the information of the registration key does not need to be included in the data record, the amount of data can be saved. For example, the katakana notation for pronunciation is prepared when the key is set, but is not actually included in the actual data, and the key identifier only has information for distinguishing it from other keys.

In the data pointer, the address that points to the beginning of the actual data record in the actual data is the delimiter F.
It is stored in 3 bytes following FH. When the data pointer is FFFFFFH, it indicates that the data record is deleted.

The actual data file is a set of data records, and the data records have the following format.

(Heading word) (Heading delimiter [NULL])
(Content part) (Record delimiter [LF]) When 80H is used as the key identifier as described above, the key is identified by referring to the entry word of this data record. However, referencing the actual data in this manner causes a decrease in speed in that the number of times the pointer is manipulated and the number of accesses to the next memory increase, so that if speed is important, the same as a normal search key is used. The key identifier may be used.

The actual data file 13 (content part) is an actual dictionary description part, but in this embodiment, since the information of the search key is not included by the key identifier, the concept of the field etc. is not necessary inside this and the contents Can be flat text. The data record becomes one text file as a whole.

Next, an algorithm for data retrieval by the English-Japanese electronic dictionary system described above will be described with reference to the flowchart of FIG.

First, after initialization (step 101), the hash value h of the search key is obtained, and the content of the position h in the hash table is read as the index position ip (step 102). Next, it is determined whether ip = FFFFFFH (step 103). Where ip = F
If it is FFFFFH, it means that the key is an unregistered key, and the process ends. If ip = FFFFFFH is not satisfied, 3 bytes are read as a chain position cp from the position ip on the chained index (step 104),
It is determined whether [ip + 3] ≧ 80H (step 1
05). Here, if a byte string of 80H or more continues from ip + 3, it is read as the key identifier kr (step 106). If omitted, the previous one is used. Next, it is determined whether the current index record corresponds to the key based on kr (step 10
7), if applicable, F on the chained index
Skip to FH and continue to 3 bytes data pointer d
It is read as p (step 108). Then, dp =
It is determined whether FFFFFFH (step 1
09. Here, if dp = FFFFFFH is not satisfied, the data record exists, so the position on the data file dp
From 0AH (=
Add up to LF to the list of results (step 11)
0). Next, it is determined whether cp = 0 (step 10
1) If cp = 0, the process ends. If not, the chain continues, so cp is substituted for ip (step 112) and the process returns to step 104.

Since the chain index and the substance of the data are separated, insertion and deletion can be realized at high speed by switching the pointers.

In the above embodiment, it is premised that the index is held in the disk device 26, but if the capacity of the main memory 28 is sufficient, the index can be held in the main memory 28 for further speeding up. it can.

The data access method according to the present invention can be applied to the chained hash method in general,
The present invention is not limited to the English-Japanese electronic dictionary system shown in the above embodiment. For example, it can be used as a basic data structure in a system for searching a large amount of information at high speed such as a text database.

[0033]

As described above, in the data access method according to the present invention, since the identifier indicating the information about the key is added to the chain pointer in the chained index, the key information for key comparison is added. Can be used efficiently, and the speed of data retrieval and the saving of data size can be achieved at the same time. As a result, it is possible to very efficiently realize a search mainly using a search key represented by an electronic dictionary.

[Brief description of drawings]

FIG. 1 is a diagram showing an outline of a data structure in an English-Japanese electronic dictionary system.

FIG. 2 is a block diagram showing a schematic configuration of an English-Japanese electronic dictionary system.

FIG. 3 is a diagram showing a structure of an index record.

FIG. 4 is a diagram showing a list of chain pointers in an index record.

FIG. 5 is a flowchart showing an algorithm for data retrieval by the English-Japanese electronic dictionary system.

[Explanation of symbols]

11 ... Hash table, 12 ... Index with chain,
13 ... Actual data file

Claims (1)

Claims: What is claimed is: 1. A data access method comprising a data structure having a hash table, a chained index, and an actual data file, wherein a chain pointer in the chained index relates to a key. A data access method characterized by adding an identifier indicating information.