JPH0561918A - Numerical retrieving device - Google Patents

Abstract

PURPOSE:To reduce the necessary capacity of a retrieving work buffer an at the same time to improve the retrieving speed. CONSTITUTION:A numeric character comparison part 13 compares the numerical character code of the text data received from a text code identifying part 12 with the numeric character code of the conditional numerical data received from a conditional numeric data reading part 18 to decide the difference of value between both data. The result of this comparison is given to a text code investigating part 10, a conditional numerical skip part 22, a conditional numerical substitution part 24, and a deciding part 20 respectively. The part 22 skips the numeric character data to be read out of a conditional numerical data holding part 16 next by an amount equal to one digit when the numeric character code is larger than the numeric character data. The part 24 substitutes the numeric character data to be read out of the part 16 next for the prescribed numerical value larger than the maximum numerical value equivalent to one digit that can be obtained out of a text when the numeric character code is different from the numeric character code. Then the part 20 decides whether each numerical data satisfies the prescribed numerical conditions or not.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a numerical search device for searching numerical data satisfying a predetermined numerical value in text data stored in a buffer memory or the like.

[0002]

2. Description of the Related Art As one mode of document search, there is a case where a numerical value satisfying a predetermined numerical value is searched. The conventional method for performing this kind of numerical value search is to search the entire text (document) in advance, and if a numerical value is detected, convert the numerical value into a binary code and set the position within the text for all numerical values. If a numerical condition is given in advance by creating a table corresponding to binary codes, the numerical condition that indicates the numerical condition and each numerical value (binary code) on the table are sequentially compared to satisfy the numerical condition. A method of determining whether or not it is a numerical value (first conventional method), a method of using an automan having a state transition path for all possible numerical values satisfying the numerical value (second conventional method), and a numerical condition for each digit There is known a method (third conventional method) in which the collation processing is performed in parallel by dividing each numerical value condition by using separate automan for each of the divided numerical conditions.

[0003]

However, each of the above conventional methods has the following drawbacks. In the first conventional method, it is necessary to search the entire text in advance and create a table in which the position in the text and the binary code are made to correspond to all numerical values, and this processing takes a considerable amount of time and overhead. Occurred.

In the second conventional method, the state transition table of Automan becomes large in scale, and an overhead occurs in the process of generating this table.

Compared to the second conventional method, the third conventional method reduces the size of the state transition table and compresses the automan, but still requires a relatively large capacity search work buffer. There was also the risk of overhead.

The present invention has been made in view of the above problems, and an object of the present invention is to provide a numerical search device capable of reducing the required capacity of the search work buffer and improving the search speed.

[0007]

In order to achieve the above object, a numerical retrieval device of the present invention comprises a conditional numerical data holding means for holding conditional numerical data indicating a predetermined numerical condition,
Conditional numerical data reading means for reading each numerical data constituting the conditional numerical data from the conditional numerical data holding means one digit at a time from the most significant digit, and text codes constituting the text data for each character in a predetermined order. Text code searching means for searching, text code identifying means for identifying whether the text code searched by the text code searching means is a numeric code, text code searching means, and identified by the text code identifying means. A numeric code comparing the numeric code read from the conditional numeric data holding means by the conditional numeric data reading means, a comparison result of the numeric comparing means, a reading result of the conditional numeric data reading means, The search result of the text code search means or the text code Depending on the identification result of the identifying means, the numerical data contained in the text data is configured to and a judgment means for judging whether the predetermined numerical condition is satisfied.

In order to further improve the search speed,
Depending on the comparison result of the numerical comparison means, the subsequent numerical data to be read from the conditional numerical data holding means is given a predetermined numerical value that is conditionally larger than the maximum possible numerical value for one digit obtained in the text. It is configured to include means for replacing

Similarly, in order to further improve the search speed, the subsequent numerical data to be read from the conditional numerical data holding means by the conditional numerical data reading means is changed in accordance with the comparison result of the numerical comparison means. Therefore, it is configured to include a unit for skipping one digit.

[0010]

[Function] A given text code contains numerical data consisting of one-digit or two-digit or more numeric codes, and the first (highest digit) numeric code of the series of numeric codes is detected by the text code search means. When searched and then identified by the text code identifying means, the conditional numerical data reading means reads the numerical data of the most significant digit of the conditional numerical data holding means. Then, the numeral comparison means compares the numeral code T of the most significant digit and the numerical data X of the most significant digit to determine the magnitude relationship between the two. next,
In the same manner as described above, the numerical code T and the numerical data X are compared with respect to each second next digit of the numerical data of the text code and the conditional numerical data.

Thus, the numerical code of each most significant digit,
Starting with the comparison of the numerical data, the numerical code of the text code and the numerical data of the conditional numerical data are compared for each digit. In such a comparison, when the number of digits of the numerical data of the text code is smaller than the number of digits of the conditional numerical data, the conditional numerical data reading unit searches by the text code searching unit before the read result indicating unreadable is output. The search result which is impossible or the identification result of the non-numeric code identification is issued by the text code identification means. On the contrary, when the number of digits of the numerical data of the text code is larger than the number of digits of the conditional numerical data, the conditional numerical data reading means outputs an unreadable read result earlier.

Therefore, the judging means determines the text according to the comparison result of the numerical comparison means, the reading result of the conditional numerical value data reading means, the search result of the text code searching means or the identification result of the text code identifying means. It can be determined whether each numerical data included in the data satisfies a predetermined numerical condition.

If the numerical comparison means outputs a comparison result (T> X or T <X) that the numerical code T of the text code and the numerical data X of the conditional numerical data do not match, the conditional numerical value is obtained. Subsequent numeric data to be read from the data holding means is of a predetermined numeric value (for example, the value "A" in 4-bit data) which is larger than the maximum possible numeric value for one digit (for example, "9"). By replacing the digit data for exception processing, the subsequent digit comparison processing can be completed simply and in a single time, and the comparison judgment processing between the number of digits of the numeric data of the text code and the number of digits of the conditional numeric data can be performed quickly. be able to.

Further, if the comparison result (T> X) that the numerical code T of the text code is larger than the numerical data X of the conditional numerical data is output from the numerical comparison means, the conditional numerical data is held. By skipping the subsequent digit data to be read from the means by one digit,
The number of digits of the condition numerical data can be known earlier, and the digit number comparison / judgment process can be performed quickly and quickly.

[0015]

Embodiments of the present invention will be described below with reference to the accompanying drawings. FIG. 1 is a block diagram showing the configuration of a numerical search device according to an embodiment of the present invention. This numerical search device
For example, one function of a word processor or personal computer is a microcomputer (CP
U, ROM, RAM, input / output device, software).

The text code search unit 10 searches the text data stored in a buffer memory (not shown) for a text code in a predetermined order (usually, an arrangement order of text characters) for each character. The text code identifying unit 12 identifies whether the text code searched (extracted) by the text code searching unit 10 is a numeric code or a non-numeric code (code such as kana, kanji, or symbol). Text code identification unit 12
The numeric code identified by is given to the numeric comparison unit 14.

The condition numerical data holding unit 16 is set with conditional numerical data indicating a numerical condition (range to be searched). In the case of operating as a stack, each numerical data constituting the conditional numerical data is written in the conditional numerical data holding unit 16 by one digit from the lowest digit. The condition numerical data reading unit 18 reads each numerical data forming the conditional numerical data from the conditional numerical data holding unit 16 one digit at a time from the most significant digit, and gives the read numerical data to the numerical comparison unit 14. The reading of the numerical data from the conditional numerical data holding unit 16 by the conditional numerical data reading unit 18 is performed at a predetermined timing, that is, in response to the numerical code being identified by the text code identifying unit 12.

The numeral comparison unit 14 compares the numeral code T of the text data received from the text code identification unit 12 with the numerical data X of the conditional numerical data received from the conditional numerical data reading unit 18, and the two of T and X are compared. Determine the size relationship. The comparison judgment result from the numeral comparing section 14 is given to the text code searching section 10, the conditional numerical value skipping section 22, the conditional numerical value replacing section 24 and the judging section 20.

The conditional numerical value skip unit 22 operates when the numerical code T of the text code is larger than the numerical data X of the conditional numerical data, and the numerical data to be read next from the conditional numerical data holding unit 16 for one digit. Make it skip. The conditional numerical value skip unit 22 operates when the numerical code T of the text data and the numerical data X of the conditional numerical data do not have the same value, and the numerical data to be read next from the conditional numerical data holding unit 16 is written in the text. Maximum number of digits that can be obtained (“9” in normal text)
A predetermined numerical value larger than that, for example, numerical data << A >> for exception processing having a value "A" in 4-bit data is replaced. The text code search unit 10 outputs the next text code after the comparison judgment is completed by the numeral comparison unit 14 and the conditional numerical value skip unit 22 or the conditional numerical value replacement unit 24 performs the conditional processing as described above. Explore.

The judging section 20 determines the text data according to the comparison result of the numeral comparing section 14, the reading result of the conditional numerical data reading section 18, the searching result of the text data searching section 10 or the identifying result of the text code identifying section 12. It is determined whether or not each of the numerical data included in the above condition satisfies a predetermined numerical condition.

Next, the operation of the numerical retrieval system of this embodiment will be described with reference to the flow chart of FIG. As an example,
For the text containing the phrase "... 57 apples, 389 tangerines, 4261 pears ..."
Numerical value search when ≦ 387 ”is given will be described.

In FIG. 2, when the numerical search is started,
In the condition numerical value data holding unit 16 as a stack memory, each numerical value “3” forming the conditional numerical value “387”,
"8" and "7" are in order from the lower digit ("7", "8",
They are written (in the order of "3") (100).

On the other hand, the text code search unit 10 accesses the buffer memory storing the text data and searches the text data one by one. The text code identifying unit 12 identifies whether the searched text data is a numeric code (102, 104).

In the case of this example, the text data << 5 >> corresponding to the numeral "5" following "Ringa ga" is a numeral code. This numeric code << 5 >> is the text code identification part 1
When it is identified by 2 and given to the number comparison unit 14,
The conditional numerical data reading unit 18 reads the character data at the top of the stack, that is, the numerical data << 3 >> of the most significant digit of the numerical condition "387" from the conditional numerical data holding unit 16,
This numerical data << 3 >> is given to the comparison unit (106, 10
8).

The numeric comparison unit 110 compares the numeric code << 5 >> from the text code identification unit 12 with the numeric data << 3 >> from the conditional numeric data reading unit 18, and the former << 5 >> is derived from the latter << 3 >>. The comparison result is also large. When such a comparison result (T> X) is issued, the conditional numerical value skipping unit 22 discards (skips) the numerical data << 8 >> at the top of the stack and outputs the numerical data << 7 >> of the next digit. Move to the numerical data at the top of the stack. In the case of this example, since the skipped numerical data exists (114), the conditional numerical substitution unit 24 replaces the numerical data << 7 >> at the top of the stack with the numerical data << A >> for exception processing (11).
6).

When the replacement process (116) is completed, the text code search unit 10 sets << 7 >> as the next text code.
(118). The text code << 7 >> is identified by the text code identification unit 10 as a numeric code (120). Then, the conditional numerical data reading unit 18 reads the numerical data << A >> of the top stack from the conditional numerical data holding unit 16, and the numerical data << A >> is read.
Is given to the numeral comparison unit 14 (106, 108, 11
0).

The numerical comparison section 14 outputs a comparison result (T <X) that the numerical data << A >> of the numerical condition is larger than the numerical code << 7 >> of the text data. Then, the process proceeds to the replacement process 116, and the numerical condition replacement unit 24
Tries to replace the numerical data at the top of the stack with the numerical data << A >> for exception processing. However, since the numerical data at the top stack no longer exists at this time, the replacement processing 116 is not actually performed. become.

Next, the text code search unit 10 searches for << piece >> as a text code (118), and this text code << piece >> is identified by the text code identification unit 10 as a non-numeric code (120). ). In response to this identification result, the determination unit 20 outputs the determination result Q that the numerical value “57” is the numerical value N that satisfies the numerical condition “N ≦ 387” (122).

Next, for each non-numeric character ", mandarin orange", the text code search and identification process (102, 104) is repeated in the text code search unit 10 and the text code identification unit 12. Then, "389
When the numerical code << 3 >> corresponding to the leading numeral "3" of "unit" is searched and identified, the condition numerical data reading unit 18
Is the character data at the top of the stack, that is, the numerical condition "38
The numerical data << 3 >> of the highest digit of "7" is read from the conditional numerical data holding section 16 and this numerical data << 3 >> is given to the comparing section (106, 108, 110).

The numeric comparison unit 110 compares the numeric code << 3 >> from the text code identification unit 12 with the numeric data << 3 >> from the conditional numeric data reading unit 18, and outputs a comparison result that both are equal. As soon as such a comparison result (T = X) is issued, the text code search unit 10 searches for the next text code << 8 >> (118). This text code << 8 >> is identified by the text code identification unit 10 as a numeric code (120). Then, the conditional numerical data reading unit 18 reads the numerical data << 8 >> of the top stack from the conditional numerical data holding unit 16 and supplies this numerical data << 8 >> to the numerical comparison unit 14 (106, 108, 110).

The numerical comparison section 14 outputs a comparison result (T = X) that both the numerical code << 8 >> of the text data and the numerical data << 8 >> of the numerical condition are equal. When this comparison result (T = X) is output, the text code search unit 10 searches again for the next text code << 9 >> (11
8), the text code identifying unit 10 identifies the text code << 9 >> as a numeric code (120). Then, the conditional numerical data reading unit 18 reads the numerical data << 7 >> of the top stack and gives this numerical data << 7 >> to the numerical comparison unit 14 (106, 108, 110).

The numeric comparison unit 110 outputs a comparison result that the numeric code << 9 >> from the text code identification unit 12 is larger than the numeric data << 7 >> from the conditional numeric data reading unit 18. When this comparison result (T> X) is issued, the conditional numerical value skip section 22 attempts to discard the numerical value data at the top of the stack (112, 114). However, in this case, since there is no leading numerical data in the stack of the conditional numerical data holding unit 16, it becomes impossible to read (11
4). This stack empty state is the condition numerical data reading unit 1
The determination unit 20 is notified by 8. Then, the determination unit 2
0 gives a determination result P that the numerical value “57” does not satisfy the numerical condition “N ≦ 387” (115, 1
24).

Then, for each non-numeric character ", pear", the text code search and identification processing (102, 104) is repeated in the text code search unit 10 and the text code identification unit 12. Then, "426
When the numerical code << 4 >> corresponding to the leading numeral "4" of "1" is searched and identified, the condition numerical data reading unit 1
8 is the character data at the top of the stack, that is, the numerical condition “3
The numeral data << 3 >> of the most significant digit of "87" is read from the conditional numerical data holding section 16 and this numerical data << 3 >> is given to the comparing section (106, 108, 110).

The numeric comparison unit 110 compares the numeric code << 4 >> from the text code identification unit 12 with the numeric data << 3 >> from the conditional numeric data reading unit 18, and the former << 4 >> is derived from the latter << 3 >>. The comparison result is also large. When such a comparison result (T> X) is issued, the conditional numerical value skipping unit 22 discards (skips) the numerical data << 8 >> at the top of the stack and outputs the numerical data << 7 >> of the next digit. Move to the numerical data at the top of the stack. In the case of this example, since the skipped numerical data exists (114), the conditional numerical substitution unit 24 replaces the numerical data << 7 >> at the top of the stack with the numerical data << A >> for exception processing (11).
6).

When the replacing process (116) is completed, the text code search unit 10 sets << 2 >> as the next text code.
(118). This text code << 2 >> is identified by the text code identification unit 10 as a numeric code (120). Then, the conditional numerical data reading unit 18 reads the numerical data << A >> of the top stack from the conditional numerical data holding unit 16, and the numerical data << A >> is read.
Is given to the numeral comparison unit 14 (106, 108, 11
0).

The numerical comparison section 14 outputs a comparison result (T <X) that the numerical data << A >> of the numerical condition is larger than the numerical code << 2 >> of the text data. Then, the process proceeds to the replacement process 116, and the numerical condition replacement unit 24
Tries to replace the numerical data at the top of the stack with the numerical data << A >> for exception processing. However, since the numerical data at the top stack no longer exists at this time, the replacement processing 116 is not actually performed. become.

Next, the text code search unit 10 searches << 1 >> as a text code (118), and the text code << 1 >> is identified by the text code identification unit 10 as a numeric code (120). In response to this identification result, the conditional numerical data reading unit 18 attempts to read the numerical data of the top stack from the conditional numerical data holding unit 16, but since the numerical data of the top stack has already been lost, it cannot be read. Become. In response to this reading result, the determination unit 20 determines that the numerical value “4261” is a numerical value that does not satisfy the numerical condition “N ≦ 387”.
Is output (109, 124).

FIG. 3 is a state transition diagram corresponding to the operation of the above-described embodiment. In this state transition diagram,
The state at the processing stage of steps 102 and 104,
2 to show states in the processing stages of steps 106 to 120 in FIG. 2, and show states in which the result P is obtained in FIG.
Further, a thin solid line vector S1 indicates a state transition in the case where a comparison judgment result (T = X) that both numbers T and X are equal is output from the number comparison process 110, and a one-dot chain line vector S1.
2 is the comparison judgment result that the numerical data X of the conditional numerical data is larger than the numerical code T of the text code (T
<X) indicates the state transition, and the thick solid line vector S3 shows the comparison judgment result (T> X) that the numerical code T of the text code is larger than the numerical data X of the conditional numerical data. The state transition in the case of being performed is shown, and the dotted vector S4 shows the state transition in the case where the one-digit number data X is read from the stack. If the text code to be searched is interrupted or terminated or a non-numeric code is identified in the middle of reaching the state, the process exits from this state transition diagram and the result Q is output ( 118, 120, 122).

The text of the above example "... 57 apples,
In the case of 389 mandarin oranges, 4261 pears ... ", when the numerical value is" 57 ", the state transitions to → and the result Q is output, and when the numerical value is" 389 "→→
→→ → state transition and result P is output, and the numerical value "326
In the case of "1", the state transitions to →→→ and the result P is output. As the number of digits of the condition numerical value increases, the state transition diagram extends in the horizontal direction and the number of transition states and the number of state transition vectors also increase.

In the above example, the case where the numerical condition “N ≦ 387” having a limit on one side is given has been described, but the numerical condition having a limit on both sides, for example, “145 ≦ N”.
When ≦ 548 is given, numerical data satisfying the numerical value condition of “145 ≦ N” and numerical data satisfying the numerical value condition of “N ≦ 548” are respectively searched, and the logical product of both search results is obtained. By taking the target numerical condition “145 ≦
A numerical value that satisfies N ≦ 548 ”can be searched.

In the above-described embodiment, the conditional numerical value skip section 22 and the conditional numerical value replacing section 24 are provided, and the stack of the conditional numerical value data holding section 16 is conditionally set according to the comparison judgment result of the numerical value comparing section 14. By skipping the leading numeral data or replacing the leading numeral data in the stack with the exception processing numeral data << A >>, the number comparison process was simplified and speeded up, and the digit number comparison process was speeded up. ,
It is also possible to omit the condition value skipping unit 22 and the condition value replacing unit 24. In that case, the numerical comparison unit 14 may accumulate the respective comparison determination results and give the comparison determination results of all digits to the determination unit 20.

[0042]

The present invention has the following effects by having the above-mentioned structure. According to the numerical search device of claim 1, the numerical code of the most significant digit and the numerical data are compared first, and then the numerical code of the text code and the numerical data of the conditional numerical data are compared for each digit and the comparison is made. Based on the result, the result of reading the condition numerical data, the result of the text code search or the result of the text code identification, it is determined whether or not each numerical data included in the text data satisfies a predetermined numerical condition. ,
There is no need to create an automan for each possible numerical data, the working buffer memory capacity is reduced, and high-speed numerical search can be performed.

According to the numerical search device of the second aspect, the subsequent numerical data to be read from the conditional numerical value holding means is conditionalized according to the comparison result of the numerical code of the text code and the numerical data of the conditional numerical data. Replace the following digit with a predetermined value that is larger than the maximum possible digit value obtained in the text code.
In a single time, the comparison / determination process of the number of digits of the numerical data of the text code and the number of digits of the condition numerical data can be performed quickly, and the numerical retrieval process can be performed at high speed.

According to the numerical retrieval device of the third aspect, the subsequent numerical data to be read from the conditional numerical data holding means is conditionalized according to the comparison result of the numerical code of the text code and the numerical data of the conditional numerical data. Therefore, by skipping by one digit, the number of digits of the conditional numerical data can be determined earlier, the comparison and determination processing of the number of digits can be performed quickly and quickly, and the numerical value retrieval processing can be performed at high speed.

[Brief description of drawings]

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

FIG. 2 is a flowchart showing an operation of the numerical value search device according to the embodiment.

FIG. 3 is a state transition diagram according to an embodiment corresponding to the flowchart of FIG.

[Explanation of symbols]

 10 Text Code Exploration Section 12 Text Code Identification Section 14 Numerical Comparison Section 16 Conditional Numerical Data Storage Section 18 Conditional Numerical Data Reading Section 20 Judgment Section 22 Conditional Numerical Skip Section 24 Conditional Numerical Replacement Section

Claims (3)

[Claims] 1. A numerical value search device for searching numerical value data satisfying a predetermined numerical value value for given text data, and conditional numerical value data holding means for holding conditional numerical value data indicating the predetermined numerical value condition, and the conditional numerical value. Conditional numerical value data reading means for reading each numerical data constituting the conditional numerical data from the data holding means one digit at a time from the most significant digit, and the text code constituting the text data is searched for one character at a time in a predetermined order. A text code search means, a text code identification means for identifying whether or not the text code searched by the text code search means is a numeric code, a text code search means searched for, and identified by the text code identification means The numeric code and the conditional numerical data reading means A numeric comparison means for comparing numeric data read from the conditional numeric data holding means, a comparison result of the numeric comparison means, a read result of the conditional numeric data reading means, a search result of the text code search means or the text Depending on the identification result of the code identification means, determination means for determining whether or not each numerical value data included in the text data satisfies the predetermined numerical value condition,
A numerical search device comprising: 2. According to the comparison result of the numerical comparison means,
It is provided with means for conditionally replacing subsequent numerical data to be read from the conditional numerical data holding means with numerical data having a predetermined numerical value larger than the maximum possible numerical value of one digit obtained in the text. The numerical search device according to claim 1, wherein 3. According to the comparison result of the numerical comparison means,
The numerical value according to claim 1 or 2, further comprising: a unit that conditionally skips one digit of subsequent numerical data to be read from the conditional numerical data holding unit by the conditional numerical data reading unit. Search device.