JPH06309369A - Information retrieving device - Google Patents

Abstract

PURPOSE:To perform efficient retrieval matching the intention of a user by handling a number part as a numeral at the time of range specified retrieval for a character string index consisting of a non-number part and a number part. CONSTITUTION:A retrieval condition analyzing means 2 analyzes a retrieval condition character string inputted from a retrieval condition input means 1, converts it into a parameter in form which can instruct retrieval processing, and sends the parameter to a retrieval process executing means 3. The means 3 operates on the basis of the parameter supplied from the means 2 to retrieve a database stored in an index information storing means 4 and a display means 6 to which a retrieval result is sent reads image data on a retrieved list out of an image information storage means 5 and displays the data. Namely, a character string showing retrieval conditions and a character string to be retrieved are processed divisionally as non-number parts and number parts respectively. Consequently, the large-small relation between the number parts can be obtained by comparison by a method different from the non-number parts as the user intends.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information retrieving apparatus for retrieving a designated range of indexes from an electronic filing system or the like, and in particular, it is constructed so that an index intended by a user can be retrieved accurately. It is a thing.

[0002]

2. Description of the Related Art In an electronic filing system, in order to identify the contents of a file, according to the user's preference,
Data is stored with an index formed of a character string of numbers or an index of a character string combining numbers and English characters. When calling a file from the electronic filing system, search this index to find the desired file.

As a method of searching for an index, there is a method of inputting two character strings and extracting all the indexes included in the range. In this range designation search, for example, the indexes are A, A1, A2, A11,
When there are ten B1, B2, B3, 1, 3, and 10, B1, B2, and B3 are to be taken out by inputting "B1" to "B3" as the search range.

In this range-specified search, first, a list of indexes to be searched, which is expressed by a character string, is sorted based on the magnitude relation of the character codes forming the index character string, and then, on the sorted list, The character code of the two character strings specified in the search range is compared with the magnitude relationship, and the character string located between the two character strings is displayed as the search result.

From the smallest character code size, "0", "1", ..., "8", "9", "A", "B", "...,"
The order is Y ',' Z ',' a ',' b ',' ... ',' y ',' z '. (Note that character codes are enclosed in single auto and written as'O '.) Also, Kanji has a character code larger than these alphanumeric characters, and symbols have a size before and after the alphanumeric character. It has a character code of Sa.

When the character strings are sorted based on the size relation of the character codes, in contrast to "left justification" in which the size relations of the character codes are compared in order from the first character of the character string, conversely,
There are two methods, "right justification", which compares the size of character codes from the character after the character string, and "Left justification" is used for the index that often adds English characters or Chinese characters at the beginning. It is common to sort.

In the "left-justified" sort, since the magnitude relations of the character codes are compared in order from the first character of the character string, a list including 10 character strings is A, A1, A2, A11, B.
When 1, B2, B3, 1, 3, 10, the sorted results are 1, 10, 3, A, A1, A11, A2, B.
The order is 1, B2, B3.

On the list sorted in this way, the magnitude relationship between the two character strings specified as the search range is compared, and the character string located between the two character strings is displayed as the result obtained. .

[0009]

As described above, in the conventional range search, since the character code is "left-justified", the numerical part included in the character string is expected to give the result expected by the user. Was not always. For example, "
Since 10 "is arranged higher than" 3 ", the search range is"
If specified as 1 "to" 3 "," 10 "is included in the search results, and if specified as" A1 "to" A2 "," A1 "and" A "
11 "and" A2 "are displayed as search results. Therefore, users were often confused by this and could not utilize the device. (Note that the character string is enclosed in double quotes, such as" ABC ". Therefore, in order to prevent this, "0", such as "001", "002", "010", etc.
There is also a method of supplementing the upper digit with aligning the number of digits (number of characters) and maintaining the magnitude relationship as a numerical value. However, this careful consideration must be taken when assigning an index in advance, which is an operational burden in creating a database that needs to store a huge amount of data.

The present invention solves these conventional problems, and provides an information retrieval device capable of accurately searching a character string by judging the magnitude relationship of numbers as intended by the user. It is intended to be provided.

[0011]

Therefore, according to the present invention, when a character string representing a search start condition and a search end condition is input, a character string included between these search conditions is searched from the search target character string list. The information retrieval device for detection has a retrieval condition analysis means for analyzing a pattern of a combination of non-numeric characters and numeric characters in a character string representing a retrieval condition, and a pattern matching the pattern analyzed by the retrieval condition analysis means. The character string extraction means that extracts each character string from the search target character string list, the size comparison means that separates the search target character string into non-numeric characters and numerical values, and the size relationship, and the character string extraction device. Sorting means for sorting each character string using the magnitude comparison means,
The search means is provided for extracting the character strings included between the search conditions from each of the character strings sorted by the sorting means by using the magnitude comparison means.

Further, the magnitude relationship comparing means is configured to treat the separated character strings of numbers as numerical values in the magnitude relationship comparison.

[0013]

In the information retrieving apparatus of the present invention, the character string representing the retrieval condition and the character string to be retrieved are separately processed into the non-numeric part and the numerical part. Therefore, it is possible to compare the magnitude relationship of the numeric part by a method different from that of the non-numeric part in a manner suitable for the user's intention.

When handling this numerical part as a numerical value,
For example, "10" is processed as a value larger than "3", so that a search matching the normal user recognition can be performed.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows, as an embodiment, a configuration in which the information retrieval apparatus of the present invention is applied to an electronic filing system.

This device comprises search condition input means (keyboard) 1 for inputting search conditions in English letters, numbers, symbols, etc.,
Search condition analysis means 2 for analyzing the input search condition character string, search processing execution means 3 for executing the search, index information storage means 4 for managing the index information of the document to be searched, and image data in page units. The image information storage means 5 managed by the user and the display means 6 for displaying the retrieved image in a form requested by the user.

The search condition analysis means 2 analyzes the search condition character string input from the search condition input means 1, converts the search condition character string into a parameter capable of instructing the search process, and transmits it to the search process execution means 3 to perform the search process. The executing means 3 operates based on the parameters given by the search condition analyzing means 2 to search the database stored in the index information storing means 4, and the display means 6 to which the search result is transmitted is searched. The image data of the list is read from the image information storage means 5 and displayed.

The operation of each means will be described in more detail.

The search condition input means 1 selects a predetermined input frame of the input panel displayed on the operation terminal with a pointing device and then inputs a search condition character string using the keyboard.

FIG. 2 shows a state of a panel for inputting search conditions at this time. In the display part of the operating terminal,
A bitmap display that can be shared with the image display means 6 is used. A search condition input panel 21 for range search is displayed on this display, and two search condition input frames 22 and 23 relating to the "image name" are provided inside the search condition input panel 21. The “image name” refers to the index information name given to one image for searching. The data type of this index information is a character string.

The input frame 22 is for inputting search start conditions, and the input frame 23
Is for inputting the search end condition, and in FIG. 2, the start condition is "
The case where A1 "and the end condition are" B10 "is shown.

Further, an "execute" button 24 and a "cancel" button 25 are displayed on the screen, and when the "execute" button 24 is instructed with a pointing device, a search is started based on the input search conditions. Therefore, the next search condition analysis means 3 is activated. Also, if you point the "Cancel" button 25 with your pointing device, this search condition input panel will appear.
Close 21 and return the screen to the previous state. It should be noted that there is no problem even if interactive input processing is performed on the character terminal without using the Gitmap display as the search condition input means 1 as in the present embodiment.

Next, the processing in the search condition analysis means 2 will be described. From the search condition input means 1, two character strings of a search start condition and a search end condition are passed as parameters.

Here, for each, it is analyzed whether or not the numeral is included, and how the numeral forms the character string together with the "non-numeric" character, and it is classified into three patterns. To do.

The first pattern is "non-numeric + numeral", which has a non-numeric character string portion at the beginning such as "AC4040", "AU807", "SA0032", etc., followed by a numeric portion. In some cases.

The second pattern is "only numbers", for example "0", "" such as "10", "607", "0078".
It is a character string composed of 10 characters of 1 ',' 2 '...' 9 '.

The third pattern is "all others", for example, "KS33A", "666", etc.

This classification processing can be easily realized by checking the character code of each character in order from the first character of the character string. As a result of the classification, the search condition pattern number 1 is stored for the first pattern, the search condition pattern number 2 is stored for the second pattern, and the search condition pattern number 3 is stored for the third pattern.

Next, the operation of the search processing execution means 3 will be described. The search processing execution means 3 performs operations as means for extracting character strings, comparing sizes, sorting, and searching. The operation in the range designation search of this search processing execution means will be described with reference to the flowchart of FIG.

First, in the primary search 31, the search condition pattern number passed from the search condition analysis means 2 is checked, and if it is 1, the index storage means 4 is referred to and the "non-display" is selected from the search index list of the database. All that correspond to the pattern of "numerical figures + numerals" are extracted together with their data IDs, and a primary search list is created.

FIG. 4 shows the structure of the search index list stored in the index information storage means 4. One data is a data ID which is a serial number in the list and an index which is a search target character string. And a pointer to an image which is a data entity.

FIG. 5 (a) is a structural diagram of the primary search list obtained by the primary search, and its data is
The index and its data ID are elements.

Following the primary search 31, a search characteristic of the present invention is carried out by sorting the index list or comparing the index character string with the search condition. The definition of the size relation of the index character strings belonging to the pattern "will be described.

First, the magnitude relation between the two character strings A and B is defined as follows. (Here, A and B are arrays in which character codes are stored in order from the first character, and represent character strings.)

First, the numbers of both characters are obtained, and the smaller one is N. Next, the character codes A (1) and B (1) of the first character of the character string are compared, and it is assumed that the larger the value, the larger the character string. If they are the same, the next character codes A (i) and B (i) of both are compared in order.
In this way, the Nth character code A (N) of the character string
And B (N) are compared, and if they are equal, the one with the larger number of characters is regarded as the larger character string. Further, at this time, when the numbers of characters of both are equal, the character strings are completely matched and the sizes of the character strings are equal. this is,
It is also a strict definition of the character string sort explained in the conventional example.

Next, the magnitude relation between the index character strings having the structure of "non-numeric characters + numerals" in the present invention is defined as follows.

When the non-numeric character parts are different, the character strings of the non-numeric character parts are compared according to the above-mentioned character string size relationship definition, and the result is regarded as the size relationship.

If the non-numeric character parts are the same, the numeric part is converted into a numerical value and compared, and the magnitude relationship is taken as the result.

In the primary search list sort 32 of FIG. 3, the primary search list is sorted on the basis of the magnitude relation of the index values described above. As a result, the primary search list is first sorted according to the size relation of the non-numeric character portion and then the numeric portion treated as a numerical value, as shown in FIG. 5 (b). At this time, since the numerical part is compared as a numerical value, for example, “A11” is positioned as being larger than “A2”.

In the subsequent first search data search 33 and final search data search 34, the search start condition and the search end condition are used to search for data belonging to the range defined by these two conditions.

In the first search data search 33, the "first search result data" is detected by the following procedure. First, the search start character string and the data of the primary search list are compared in order from the first data of the list, and whether the search start condition is equal to the search start condition,
The first data that is larger than the search start condition is detected.
Next, the detected index value is compared with the search end condition, and if it is equal to or smaller than the condition value,
Set as search result data. If such an index is not detected up to the end of the list, the number of hits in the search is set to 0 and the process ends.

Next, in the final search data search 34, "final search result data" is detected by the following procedure. First, the search end condition and the index value of the data of the primary search list are compared in order from the data next to the first search result data obtained by the first search data search 33, and the index value is the first search end condition. Ask for bigger data than. The data immediately before this data, that is, the next largest data after this data is the final search result data. If such data is not detected up to the end of the list, the final data of the list will be the final search result data.

The first search result data and the final search result data obtained by the first search data search 33 and the final search data search 34 may coincide with each other.

Next, the search result list creation 35 creates the search result as its data ID list. That is, the data IDs of all the data located between the first search result data and the final search result data obtained by the first search data search 33 and the final search data search 34 are extracted, and
A search result ID list as shown in (c) is created and passed to the display means 6.

The display means 6 displays the retrieved image data
For example, the search result ID list is continuously displayed from the top member. For that purpose, the data ID is converted into a pointer representing the location of the data entity using the index information storage means 4. Specifically, the correspondence between the data ID and the pointer is read in the index information list of FIG.

After obtaining the pointer to the image data entity, the pointer to the image is passed to the image information storage means 5 to access it, and the desired image data is read out and sent to the display means 6 to display the image. The display method of the search result is
Various things can be considered depending on the purpose of the system and the user.

If the search condition pattern number passed from the search condition analysis means 2 to the search processing execution means 3 is “only numbers”, the search processing execution means 3 will be provided with a search start numerical value and a search end as parameters. Only numbers are given.

At this time, the search processing execution means 3 first takes out a character string consisting of only numbers from the index information list and creates a primary search list. Next, the index character string is treated as a numerical value, this list is sorted, and those that are equal to or higher than the search start numerical value and equal to or lower than the search end numerical value are stored in the search result ID list.

When the search character string pattern number is "other", the range designation search may be performed by the conventional method.

[0050]

As is apparent from the above description of the embodiments, in the information retrieval apparatus of the present invention, the numeric part is treated as a numeric value in the range designation search for the character string index composed of the non-numeric part and the numeric part. Therefore, an efficient search suitable for the user's intention can be realized.

[Brief description of drawings]

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

FIG. 2 is an input screen of search condition input means in the apparatus of the embodiment,

FIG. 3 is a flowchart showing the operation of the search processing execution means of the embodiment.

FIG. 4 is a structure of data stored in an index information storage unit of the embodiment,

FIG. 5 is a diagram showing a structure of data stored during a search by a search process executing means of the embodiment.

[Explanation of symbols]

 1 search condition input means 2 search condition analysis means 3 search processing execution means 4 index information storage means 5 image information storage means 6 image display means 21 search condition input panel 22 search start condition input frame 23 search end condition input frame 24 execute button 25 Cancel button

Claims (2)

[Claims] 1. An information search apparatus for detecting a character string included between these search conditions from a search target character string list when a character string representing a search start condition and a search end condition is input. Search condition analyzing means for analyzing a pattern of a combination of non-numeric characters and numeric characters in a character string representing, and each character string having a pattern matching the pattern analyzed by the search condition analyzing means is the search target character string. A character string extracting means for extracting from the list, a size comparing means for separating the search target character string into non-numeric characters and numbers and comparing the size relationship, and each character string extracted by the character string extracting means for the size comparison. Sorting means for sorting using the means, and the character strings included between the search conditions from the respective character strings sorted by the sorting means. Information retrieval apparatus characterized by providing a search means for extracting with comparison means. 2. The information retrieving apparatus according to claim 1, wherein the magnitude relation comparing means treats the separated character strings of the numbers as numerical values in the magnitude relation comparison.