JPS6243730A - Information retrieval device - Google Patents

Abstract

PURPOSE:To execute a high-speed retrieval by providing an information control means, a means indexing equivalent plural attributes on a record, and a means retrieving all the indexed attributes. CONSTITUTION:The value '0' in an equivalent attribute number column 84 indicates that its attribute has no index and the value '0' in the equivalent attribute number column 85 indicates that there is no opposite party regarded as the equivalent attribute for the retrieval operation. When an indication to index any of the attributes 1(M+1)-1(M+N) is given to a record 10 constituted in such a way, the indexing means 2 identifies the attributes 1(M+1)-1(M+N) as equivalent attributes in terms of the indexing by referring to the equivalent attribute number column 84 related to indexing in the record control information 8, and automatically makes all the attribute values in the attributes 1(M+1)-1(M+N) correspond to the serial record numbers #100s to register them into the index 70. Thereby the quick retrieval can be expected.

Description

DETAILED DESCRIPTION OF THE INVENTION FIELD OF INDUSTRIAL APPLICATION The present invention relates to an information retrieval device generally referred to as a database.

Conventional VL Technique In conventional information retrieval devices, it is not allowed for a certain attribute in a record to have multiple values;
As shown in Figure 1, the attribute 503 has attribute values KW+ ~KWN.
For records in situations where the
It was transformed as shown in Figure (a) or Figure 12 (b), and had a hierarchical or relational structure.

Problems to be Solved by the Invention However, in the configuration shown in FIG. 12(a), the structure of the record is greatly modified, making handling complicated. Even with a configuration like this, the structure of the record was considerably modified, and there was a serious problem of duplication of information.

Now, if the maximum value of the number of repeated attribute values (the number of values having a plurality of values) of the attribute 503 shown in FIG. 11 is determined to be N, a configuration as shown in FIG. 13 could also be considered.

However, in this configuration, the attributes 5031 to 5 of the repeat part
Since 03N is independently independent, the search condition expression becomes quite complicated when searching for the attribute value included in any of the attributes, resulting in a significant decrease in search speed.

In view of this point, the present invention aims to provide high-speed detection using the configuration shown in FIG.

An information retrieval device according to the present invention has an information management means for managing information at an intermediate level such as a record consisting of a plurality of attributes, and a plurality of equivalent attributes in a seed. An index creation method that creates a common index for a group of attributes in a record. In the present invention, with the above-mentioned configuration, the index creation means is equipped with a search means for searching records managed by the information management means. A common index is automatically created for a plurality of equivalent attributes, and a common index is automatically created for a plurality of equivalent attributes Pt for a search in the first search stage (which is managed by the information management means). By automatically searching all two or more indexes, attribute values across a plurality of attributes t1 in the [T]-word can be searched at high speed.

Embodiment FIG. 1 shows a block diagram of an information retrieval device d in one example of the present defense. In FIG. 1, 1 is an information management means, 2 is an index creation means, and 3 is a search means, and 4 is an index.

The information management means 1 manages information consisting of a collection of a plurality of attributes in units called records, and manages the collection of records on the main storage or secondary storage.

(If we consider the case where records are of fixed length and all records are held in secondary storage, the storage format is
It will look like the figure. In Figure 2, 5 is secondary memory, 51
~5n is a record. All records are given a one-note serial number, and the storage location of each record is (
You can find out by looking at the record serial number x (record length).

Since a record consists of a plurality of attributes, there is information for managing the format of the record, an example of which is shown in FIG. In Figure 3, 6 is a field for record definition information, 6o is a field for attribute number, 61 is a field for attribute name, 62 is a field for attribute type (character string, numerical value, etc.), and 63 is a field for attribute length and width. Column 64 is an equivalent attribute number related to index creation, and 65 is a column indicating an equivalent attribute number related to search, that is, an attribute number that should be referred to at the same time when searching by attribute values.

When the index creation means 2 is instructed to create an index regarding a certain attribute X in a record, the index creation means 2 creates the record definition information 6
By referring to , the position of attribute X in the record and its length are known, and index 4, which is the correspondence information between the value of the attribute and the record serial number, is created (FIG. 4). In Figure 4, 4
1 is the attribute value, 42 is the name of the record with that attribute value] -
This is the serial number.

The search means 3 compares the attribute value 41 in the index 4 with the attribute value given as the search condition (match, mismatch, size, inclusion, etc.), and when it finds an attribute 1i[41 that matches the condition, it corresponds to that attribute value. A record serial number 42 is obtained, and a desired record is retrieved from the secondary storage 5 using the record serial number 42.

The operation of the information retrieval device of this embodiment configured as described above will be described below.

FIG. 5 shows a first form of management of records (FIG. 13) by the information management means 1 in this embodiment. In FIG. 5, 10 is a record, and 11 to I
M, 1(M+1) to 1(M+N) are its attributes.

Here, 1(M+1) to 1(M+N) are assumed to be the first attribute group with respect to index creation.Therefore, the record management information in this example is the 6th attribute group according to the definition in FIG.
It will look like the figure. 53 in Figure 6, equal attribute number column 8
The value ○ in 4 indicates that the attribute does not have an index, and the value 0 in the equal attribute number column 85 indicates that there is no partner that is taken as an equivalent attribute during a search.

Attributes 1 (M+1) to 1 (
When an instruction is given to create an index with respect to any one of M+N), the index creation means 2 creates the record management information 8.
By referring to the equivalent attribute number column 84 for index creation, it is identified that attributes 1 (M+1) to 1 (M→N) are equivalent attributes for index creation, and the above-mentioned enlightenment automatically Attribute 1 (M → -1) ~ 1 (M + N
) and the record serial number #100 are registered in the index 70 in association with each other (FIG. 7).

Similarly, the search means 3 can determine that attributes 1(M+1) to 1(M+N) have one common index 70 by referring to the equal attribute number column 84 related to index creation in the record management information 8. By identifying the index 70 and searching the index 70, it is possible to find the record 10, regardless of whether the search target for a given attribute value is any of the attributes +(M-+-1) to +(M+N). Become.

FIG. 8 shows records by information management means 1 (FIG. 13)
This represents the second management form.

In FIG. 8, 20 is a record, 2L-2M,
2(M+1) to 2(M+N) are attributes. Here 2
(M+1'') to 2(M+N) are the equivalent second attribute group for search. Therefore, the record management information is the 9th attribute group.
It will look like the figure. In Fig. 9, as in Fig. 6, etc. (I
A value O in the attribute number column 94 indicates that the attribute does not have an index, and a value O in the equal attribute number column 95 indicates that there is no other attribute that can be treated as an equivalent attribute during a search.

Attributes 2 (M+1) to 2 (
When an instruction is given to create an index regarding attributes 2(M+1) to 2(M+N), the index creating means 2 refers to the attribute number 94 related to index creation in the record management information 9. N) does not have equal attributes with respect to index creation, and creates independent indexes 71 to IN for each attribute of attributes 2(M+1) to 2(M+N) one by one using the above-mentioned function. (Figure 10).

The search means 3 retrieves the attribute 2 (M+1
) to 2(M+N) are attributes that are not equivalent with respect to the search, that is, they should be searched at the same time), and the search target for the given attribute value is attribute 2(M+1)
Attribute 2 (M+1) for any of ~2 (M + N)
By automatically searching all the indexes 71-7N for ~2(M+N), it becomes possible to locate the record 20.

Here, the above-mentioned operation will be explained according to a flowchart.

(A) Operation during index creation FIG. 14 is a flowchart illustrating the index creation operation.

(INDI)...Start of index creation.

(IND2>...Record management information 8 or 9
Obtain the attribute configuration (=FORM) of the target record and substitute the number of attributes to variable N.

(IND3)...Start processing from the first attribute of the obtained FORM (i=1).

(IND4)...Substitute the value of equality attribute number a4 or 94 related to index creation to variable I.

(IND5)...If l-0 (no index) (
Proceed to IND8).

(IND6)...Substitute the first attribute value of the record to variable V.

(lND7)... Attribute mv is paired with the record serial number and registered in index I.

(IND8)... Increase 1 by 1.

(IND9)...If i≦N, proceed to (IND4).

(INDIO)...End of index creation.

In this way, the equivalent attribute group related to index creation is a physical combination, and the index is automatically updated and created each time a new record is registered, the contents of an existing record are updated, or a new index creation is specified.

(a) Operation at the time of search FIGS. 15A and 15B together are a flowchart for explaining the search operation in the present invention.

(FND　1)・・・・・・Start of search.

(FND2)... According to record management information 8 or 9, the attribute configuration (-FORM) of the target record is set to 19, and the number of attributes is assigned to variable N. Also, clear the searched index number storage array and search result storage array R.

(FND3)...Substitute the value of equal attribute number 85 or 95 related to the search for the attribute to be searched into variable @K.

(FND4)...If K-0 (no equivalence index), proceed to (FND16).

(FND5)...Start processing from the first attribute of the obtained FORM (i=1).

(FND6)...The first attribute t't in variable J
Assign the value of equality attribute number 85 or [θ 95 related to the search for .

(FND7)...If JfK (FND7)...
Proceed to 13>.

(FND8)...Related to the index creation of the first attribute, etc. (6 Assign the value of attribute number 84 or 94 to variable 1.

(FND9)...If 1=o (no index) then (
Proceed to FND13).

(F N D 10)...I is already 6 in the array
If it is recorded, proceed to (FND 13).

(FND1+)... Add I to the array.

(FND12)... Perform a search on index I and add the results to array R.

(FND13)... Increase 1 by 1.

(F N D 14)...If 1≦N, (F N
Proceed to D6).

(FNDI5)... Proceed to (FND19).

(F N D 16)・・・・・・Etc. 1iIIIi Attribute number 8 related to creating an index of the attribute to be searched for in variable I
Assign a value of 4 or 94.

(FND 17)...l-0 (no index)
If so, proceed to (FND 19).

(FND 18)... Perform a search on index I and add the results to array R.

(F　N　D　19)...End of search.

In this way, the search-related attribute groups are logical combinations, and by specifying them before a search operation, flexible searches can be performed without physically reorganizing the index structure.

As described above, according to this embodiment, by defining an equivalent attribute group related to index creation and an equivalent attribute group related to search for multiple attributes in a record, attributes that span multiple attributes in a record are defined. Value searches can be performed quickly.

For information on how to create an index, please refer to ISAM, B-TRE
Although there are several methods such as E and HASH, the present invention is not limited to any of these methods.

In addition, equivalent attribute groups related to index creation and search related (
It can be defined independently of the iIli attribute group.

In the present invention, as described above, the equivalent attribute number field for index creation and/or the equivalent attribute number field for search is provided. Values can be handled as if they were contained within a single attribute. That is, the combination of attributes to be searched can be freely determined and changed after the data structure is created.

As described in detail, according to the present invention, when information is managed by expanding an attribute having a maximum of N repeated attribute values into N attributes having at most one attribute value, , Can I also use a search condition expression to search for an attribute value included in any of the N attributes corresponding to the repeat part? ! It can be expected to perform searches at high speed without complication, and its practical effects are great.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an information retrieval device according to an embodiment of the present invention, FIG. 2 is a diagram showing the storage format of records in secondary storage, and FIG. 3 is a diagram showing the structure of record definition information.
Figure 4 shows the structure of the index] Figures 1 and 5 are conceptual diagrams showing the first information holding method in the same embodiment, and Figure 6 is a diagram showing record definition information for the first information holding method. , FIG. 7 is a diagram illustrating registration in the index for the first information retention method, FIG. 8 is a conceptual diagram showing the second information retention method in the same embodiment, and FIG. 9 is a diagram illustrating the second information retention method. FIG. 10 is a diagram showing the record definition information for the second information retention method, FIG.
1 is a diagram conceptually showing the structure of records handled by the apparatus of the present invention, and FIGS. 12 and 13 are conceptual diagrams representing information retention methods in conventional information retrieval [j]. FIG. 14 is a flowchart explaining the index creation operation in the same embodiment, and FIGS. 15A and 153 are flowcharts explaining the search operation. 1... Information management means 2... Index creation means 3...
・Search lower row 4...Index 5...Secondary memory 51-5
n...Record 6...Record fixed 6 information 40
．． 20...Record 11~i(M+N>,
21~2 (M -+-N)... Attribute 70.71~7N... Index 64,84.94...
・Isodental number column for index creation, 65, 85.9
5... Equivalent attribute number column related to search Patent applicant Matsushita Electric Industrial Co., Ltd. Figure 1 Figure 2 Figure 3 Figure 4 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 14 Figure 15A Figure 15B

Claims (2)

[Claims] (1) Information management means for managing information in units of records consisting of a plurality of attributes; index creation means for creating an index of attribute values for one or more specific attributes in the record; a search means for obtaining the target record by searching, and the index creation means identifies that a plurality of attributes in the record have equivalent meanings, and identifies each attribute in the equivalent attributes. A common index is created for the values, and the search means, like the index creation means, identifies that the plurality of attributes in the record have equivalent meanings, and searches for each attribute value. An information search device characterized by having a function of searching for the target record by referring to one or more of the created indexes. (2) The search means identifies that multiple attributes in the record have equivalent meanings, and refers to all indexes that have been independently created for the multiple attributes that have equivalent meanings. The information search device according to claim 1, characterized by having a function of searching the record.