JP6007221B2 - Abnormal test data generation apparatus, method, and program for multiple input variables - Google Patents

Description

  The present invention relates to an abnormal test data generation apparatus, method, and program for a plurality of input variables.

  2. Description of the Related Art Conventionally, there is a technique for automatically generating test data input to a program input variable for a program black box test based on software design (see, for example, Patent Document 1 and Patent Document 2).

  By using such a technique, it is possible to derive test data that is effective and efficient for testing the program.

JP 2010-267023 A JP 2012-181782 A JP 2012-221313 A

  However, conventionally, it has been difficult to generate efficient and effective test data for a plurality of input variables including two or more input variables having a dependency relationship with each other.

  Here, the input variable dependency relationship refers to a relationship in which the value of one input variable depends on the value of the other input variable in a constraint to be satisfied by the input variable. An example of such a constraint is a constraint such that X + Y ≦ 100 when the input variables are X and Y. In this case, in order to satisfy the restriction, the value of the input variable X and the value of the input variable Y are mutually dependent.

  Then, in the test to confirm that the input check works correctly when abnormal test data is input to the program to be tested, it is appropriate to determine which variable the bug originates from when a bug occurs. Generation of abnormal test data is important to enable identification.

  The present invention has been made in view of the above points, and an object of the present invention is to generate abnormal test data relating to a plurality of input variables including input variables having a dependency relationship with each other.

Therefore, in order to solve the above-described problem, the disclosed test data generation device determines whether or not there is a dependency between the input variables based on one or more conditional expressions to be satisfied by a plurality of input variables in the test target program, Classifying the plurality of input variables into a plurality of groups based on the presence or absence of the dependency relationship, a variable classifying unit that associates each conditional expression with a group to which the input variable related to the conditional expression belongs, and for each group, Based on the conditional expression associated with the group, a value candidate generating unit that generates a candidate for the value of the input variable belonging to the group, and for each conditional expression, the first input variable associated with the conditional expression belongs For each candidate that does not satisfy any of the conditional expressions associated with the first group, the first group A combination generation unit that selects the candidates satisfying all the conditional expressions associated with the second group one by one from each of the second groups other than the group, and generates a combination of the candidates. The candidate that the combination generation unit selects from the second group for one conditional expression is common to all the combinations generated for the conditional expression.

  It is possible to generate abnormal test data relating to a plurality of input variables including input variables having mutually dependent relationships.

It is a figure which shows the structural example of the test data generation system in embodiment of this invention. It is a figure which shows the hardware structural example of the test data generation apparatus in embodiment of this invention. It is a figure which shows the function structural example of the test data generation apparatus in embodiment of this invention. It is a flowchart for demonstrating an example of the process sequence which a test data generation apparatus performs. It is a figure which shows an example of the definition of the data structure of a design model. It is a figure which shows an example of a design model. It is a figure which shows the example of extraction of a test item list. It is a figure which shows an example of the test data generation conditions regarding one test item. It is a figure which shows the example by which the test data generation conditions were distributed to each variable set. It is a figure which shows the example of a production | generation of the value candidate for every variable set. It is a figure which shows an example of a test model. It is a figure which shows an example of the definition of the data structure of a test model. It is a flowchart for demonstrating an example of the process sequence of the classification process of an input variable. It is a flowchart for demonstrating an example of the process sequence of the production | generation process of the value candidate using a domain test technique. It is a flowchart for demonstrating an example of the process sequence of a value candidate production | generation process. It is a flowchart for demonstrating an example of the process sequence of the production | generation process of an in point value candidate. It is a flowchart for demonstrating an example of the process sequence of the production | generation process of an out point value candidate. It is a flowchart for demonstrating an example of the process sequence of the production | generation process of a test case.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram illustrating a configuration example of a test data generation system according to an embodiment of the present invention. In FIG. 1, a test data generation device 10 is communicably connected to one or more user terminals 20 via a network such as a LAN (Local Area Network) or the Internet.

  The test data generation device 10 is a computer that generates test data for input with respect to an input variable of a program used for a black box test of the program. The user terminal 20 is a terminal that functions as a user interface for the test data generation device 10. Note that the test data generation device 10 may include a display device such as a display, and input devices such as a mouse and a keyboard, and may be directly operable by the user.

  FIG. 2 is a diagram illustrating a hardware configuration example of the test data generation apparatus according to the embodiment of the present invention. The test data generation device 10 in FIG. 2 includes a drive device 100, an auxiliary storage device 102, a memory device 103, a CPU 104, an interface device 105, and the like that are mutually connected by a bus B.

  A program for realizing processing in the test data generation apparatus 10 is provided by a recording medium 101 such as a CD-ROM. When the recording medium 101 storing the program is set in the drive device 100, the program is installed from the recording medium 101 to the auxiliary storage device 102 via the drive device 100. However, the program need not be installed from the recording medium 101 and may be downloaded from another computer via a network. The auxiliary storage device 102 stores the installed program and also stores necessary files and data.

  The memory device 103 reads the program from the auxiliary storage device 102 and stores it when there is an instruction to start the program. The CPU 104 executes functions related to the test data generation device 10 in accordance with a program stored in the memory device 103. The interface device 105 is used as an interface for connecting to a network.

  FIG. 3 is a diagram illustrating a functional configuration example of the test data generation apparatus according to the embodiment of the present invention. In FIG. 3, the test data generation apparatus 10 includes a design model extraction unit 11, a design model analysis unit 12, a test item extraction unit 13, a condition extraction unit 14, a variable classification unit 15, a value candidate generation unit 16, and a combination generation unit 17. Etc. Each of these units is realized by processing that one or more programs installed in the test data generation apparatus 10 cause the CPU 104 to execute. The test data generation device 10 also uses the value candidate storage unit 111. The value candidate storage unit 111 can be realized using the memory device 103 or the auxiliary storage device 102, or a storage device connected to the test data generation device 10 via a network.

  The design model extraction unit 11 receives design document data from the user terminal 20 and extracts a design model from the design document data. Design document data refers to data indicating a software design document. The design model refers to information related to software specifications. For example, the design model includes constraints on input variables. The design model analysis unit 12 analyzes the design model and extracts each of one or more business logics included in the design model as a test target. The business logic is, for example, one function in a function group included in software, a concept corresponding to one processing procedure among a plurality of processing procedures executed by the software on the computer, or execution of the processing procedure on the computer. A concept corresponding to the program to be executed. In the design model, for each business logic, input variables for the business logic, constraints for each input variable, and the like are defined.

  The test item extraction unit 13 extracts a list of test items (hereinafter referred to as “test item list”) from each test target (each business logic). The list of test items is extracted based on restrictions on input variables included in each test target. The condition extraction unit 14 extracts, for each test item, a conditional expression (hereinafter referred to as “test data generation condition”) for generating test data for the input variable focused on in the test item from the design model.

  For each test item, the variable classification unit 15 analyzes the test data generation condition regarding the test item, and classifies the input variable regarding the test item based on the presence or absence of the dependency. A group generated as a result of classification is hereinafter referred to as a “variable set”. The variable classification unit 15 also assigns (associates) each test data generation condition to a variable set to which an input variable related to the test data generation condition belongs.

  The value candidate generator 16 for each variable set, based on the test data generation conditions assigned to the variable set, test data candidates of the input variables belonging to the variable set (hereinafter referred to as “value candidates”). Is generated. The value candidate generation unit 16 switches the method of generating value candidates based on the characteristics of the variable set (whether there is a dependency relationship between input variables included in the variable set). The value candidate storage unit 111 stores the value candidates generated by the value candidate generation unit 16.

  The combination generation unit 17 generates one or more combinations as test data for combinations of value candidates for each input variable. The combination generation unit 17 generates a test model including the generated test data, and outputs the content of the test model to the user terminal 20 as a test case table.

  Hereinafter, a processing procedure executed by the test data generation apparatus 10 will be described. FIG. 4 is a flowchart for explaining an example of a processing procedure executed by the test data generation apparatus.

  In response to an instruction input by the user to the user terminal 20, the user terminal 20 transmits software design document data to the test data generation device 10. The design document data is received by the design model extraction unit 11 in the test data generation apparatus 10. The design model extraction unit 11 analyzes the design document data based on the definition of the data structure of the design model as shown in FIG. 5, and extracts the design model as shown in FIG. 6 (S100).

  FIG. 5 is a diagram illustrating an example of the definition of the data structure of the design model. In FIG. 5, the data structure of the design model is defined by BNF (Backus-Naur Form).

  According to FIG. 5, in the present embodiment, the design model includes one or more business logics. The business logic includes a business logic ID, one or more input variables, one or more input constraints, one or more domain definitions, and the like. An input variable is a variable input to business logic. An input constraint is constraint information related to an input variable. A domain definition is a definition of a condition that an input variable must satisfy in order to cause the behavior for each business logic behavior (for each output result).

  Note that descriptions of input variables, input constraints, and structures below the domain definition are omitted.

  FIG. 6 is a diagram illustrating an example of a design model. As shown in FIG. 6, in this embodiment, the design model includes two business logics. One is the “judgment-type insurance enrollment determination process”, and the other is “the couple-type insurance charge calculation process”. The “judgment type insurance enrollment determination process” is a process (business logic) for determining whether or not the couple type insurance can be enrolled. “Couple-type insurance fee calculation processing” is a process (business logic) that calculates the fee for couple-type insurance to be enrolled when it is determined that “joint-type insurance subscription determination processing” is possible. is there. In the present embodiment, the description of the test data generation related to the “couple-type insurance fee calculation process” is omitted for the sake of convenience.

  There are four input variables with the husband ID, wife age, plan name, or remarks as variable IDs in the “coupled insurance enrollment determination process”. Husband age is an integer type variable into which the age of the husband is input. The wife age is an integer type input variable into which the wife's age is input. The plan name is an enumerated input variable into which a plan name of a plan to be subscribed among a plurality of plans belonging to the couple-type insurance is input. The comment is a character string type variable into which an arbitrary character string is input. Note that the husband age, wife age, and plan name must be entered. On the other hand, input is not required for remarks.

  As the input constraints, there are four input constraints having the constraint (1), the constraint (2), the constraint (3), and the constraint (4) as input constraint IDs, respectively. Constraint (1) includes equation (1). Formula (1) indicates that the husband's age is 18 years or older. Constraint (2) includes equation (2). Equation (2) indicates that the wife is 16 years old or older. If at least one of the constraint (1) and the constraint (2) is violated, an error message “age invalid” is displayed. Constraint (3) includes equation (3). Equation (3) indicates that the remark length is 10 characters or less. Constraint (4) includes equation (4). Formula (4) shows that each character of the remarks is a full-width kanji or a full-width katakana. When at least one of constraint (3) and constraint (4) is violated, an error message “remarks illegal” is displayed.

  As domain definitions, there are two domain definitions in which domain 1 and domain 2 are domain IDs, respectively. Domain 1 includes formula (5). Equation (5) indicates that the difference between the husband's age and the wife's age is 20 years old or less. If the expression (5) is satisfied, it is determined that the subscription is possible, and a button for starting the “couple calculation insurance fee calculation process” is displayed. Domain 2 includes formula (6). Equation (6) shows that the difference between the husband's age and the wife's age is over 20 years. If equation (6) is satisfied, it is determined that the subscription is not possible, and a link to personal insurance is displayed.

  As is clear from the above formulas (5) and (6), in the present embodiment, the husband age and the wife age are examples of input variables having a dependency relationship with each other. That is, in order to satisfy Expression (5) or Expression (6), the value of the husband age and the value of the wife age depend on the other value.

  The design model may be given by tabular data as disclosed in Patent Document 2, for example.

  Subsequently, the design model analysis unit 12 analyzes the design model and extracts each business logic included in the design model as a test target (S110). That is, in the present embodiment, one business logic is one test target. Subsequent steps S120 to S180 are executed for each test target (for each business logic). Hereinafter, a test target that is a processing target is referred to as a “target test target”.

  In step S120, the test item extraction unit 13 extracts a test item list from the target test target. Specifically, the test item extraction unit 13 generates test items for each input constraint and each domain definition included in the target test target. For example, when the target test target includes n input constraints and m domain definitions, an + bm test items are extracted. The coefficients a and b indicate that a plurality of test items may be generated from a plurality of viewpoints regarding one design element (such as an input constraint or a domain definition). In this embodiment, it is assumed that a = 1 and b = 1.

  FIG. 7 is a diagram illustrating an example of extracting a test item list. FIG. 7 shows an example in which a test item list including six test items having test item IDs 1 to 6 is extracted. Each test item includes a test item ID, a target design element, test item contents, and the like. The test item ID is identification information of each test item. The target design element is a design element (input constraint or domain definition) from which the test item is extracted. The test item content is information indicating the content of the test item.

  In FIG. 7, the test item whose test item ID is 1 (hereinafter referred to as “test item 1”, and other test items follow the same naming convention) is the constraint of the input constraints of the design model shown in FIG. This is a test item extracted based on (1). Test item 2, test item 3, and test item 4 are test items extracted based on constraint (2), constraint (3), and constraint (4), respectively. Test item 5 is a test item extracted based on domain 1 of the domain definition. The test item 6 is a test item extracted based on the domain 2.

  Subsequent steps S130 to S180 are executed for each test item extracted in step S120. Hereinafter, the test item to be processed is referred to as “target test item”.

  In step S130, the condition extraction unit 14 extracts a test data generation condition (hereinafter referred to as “test data generation condition”) corresponding to the test item of interest (S130). Specifically, when the target test item is one of the test items 1 to 4 extracted based on the input constraint, in addition to the input constraint of the target test item, all other input constraint expressions are also acquired. Test data generation conditions are generated. Therefore, for the test items 1 to 4, test data generation conditions including the respective expressions of constraint (1) to constraint (4) are generated. On the other hand, when the test item of interest is the test item 5 or 6 extracted based on the domain definition, in addition to the domain definition of interest in the test item, all input constraint expressions are also acquired, and test data is generated. A condition is generated. Therefore, for example, for the test item 5, test data generation conditions as shown in FIG. 8 are generated.

  FIG. 8 is a diagram illustrating an example of test data generation conditions for one test item. As shown in FIG. 8, for the test item 5, five test data generation conditions of Formula (1) to Formula (5) are generated. Expressions (1) to (4) are test data generation conditions for constraints (1) to (4), and Expression (5) is a test data generation condition for domain 1.

  For the test item 6, a test data generation condition including Expression (6) regarding the domain 2 is generated instead of Expression (5). The test data generation condition for each of the input constraints for the test items based on the domain definition is generated because the test for the domain definition cannot be executed correctly unless all the input constraints are satisfied. .

  Subsequently, the variable classifying unit 15 analyzes each test data generation condition, classifies the input variable related to each test data generation condition into a variable set based on the presence or absence of dependency, and sets each test data generation condition to each variable set. (S140).

  FIG. 9 is a diagram illustrating an example in which the test data generation conditions are distributed to each variable set. FIG. 9 shows an example in which the test data generation conditions shown in FIG. 8 are distributed to three variable sets 1 to 3.

  The variable set 1 is a variable set including husband age and wife age. That is, since husband age and wife age have a dependency relationship with each other, they are included in the same variable set. The variable set 2 is a variable set including a plan name. The variable set 3 is a variable set including remarks. Since each of the plan name and remarks has no dependency relationship with other input variables, a variable set is formed independently.

  Each variable set is assigned a test data generation condition including an expression relating to an input variable included in the variable set. Therefore, Expression (1), Expression (2), and Expression (5) are assigned to the variable set 1, and Expression (3) and Expression (4) are assigned to the variable set 3. Since there is no formula related to the plan name, there is no test data generation condition assigned to the variable set 2.

  Subsequent steps S150 to S170 are executed for each variable set. Hereinafter, the variable set to be processed is referred to as a “target variable set”.

  In step S150, the value candidate generator 16 determines whether or not two or more input variables are included in the target variable set. For example, if the target variable set is the variable set 1, it is determined that two or more input variables are included. If the target variable set is the variable set 2 or 3, it is determined that two or more input variables are not included.

  When two or more variable sets are not included in the target variable set (NO in S150), the value candidate generating unit 16 selects a value candidate related to one input variable included in the target variable set, for example, Patent Document 1 or Patent It produces | generates using the method described in the literature 2 (S160). The value candidate generation unit 16 stores the generated value candidate in the value candidate storage unit 111 in association with information indicating the feature of the value candidate.

  On the other hand, if the target variable set includes two or more variable sets (YES in S150), the value candidate generating unit 16 generates a value candidate for the input variable included in the target variable set using the domain test technique. (S170). The value candidate generation unit 16 stores the generated value candidate in the value candidate storage unit 111 in association with information indicating the feature of the value candidate. Details of step S170 will be described later.

  When steps S150 to S170 are executed for the variable sets 1 to 3, the value candidate storage unit 111 stores value candidates as shown in FIG.

  FIG. 10 is a diagram illustrating an example of generating value candidates for each variable set. FIG. 10 shows seven values of an in point, an out point, an on point, two normal value off points, and two abnormal value off points with respect to a set of husband age and wife age included in the variable set 1. Candidates are generated, three value candidates including two normal values and one abnormal value are generated for the plan names included in the variable set 2, and four normal values and four values for the remarks included in the variable set 3 are generated. An example in which eight kinds of value candidates including abnormal values are generated is shown.

  In the present embodiment, for the variable set 2 and the variable set 3, normal values and abnormal values are generated as value candidates in step S160. "Abnormal value" is included. For variable set 1, normal values and abnormal values are generated as value candidates in step S <b> 170, and therefore, the characteristics of these input variable value candidates include “normal value” or “abnormal value”. Yes.

  Further, regarding the value candidates of the variable set 1, (in point), (out point), (on point), or (off point) are included in the feature. “On point” indicates that the value candidate is an on point. The “on point” is a value on the boundary of the equivalence class obtained based on the input constraint or the domain definition. “Out point” indicates that the value candidate is an out point. The out point is a value outside the equivalence class. “In point” indicates that the value candidate is an in point. The in point is a value in the equivalence class. The “off point” indicates that the value candidate is an off point. The off point is a value in the vicinity of the boundary obtained by shifting the boundary where the on point is placed to the plus side and the minus side for each variable. However, when there is only one on point as shown in Expression (5), a value shifted by 1 from the on point to each of the plus side and the minus side becomes the off point.

  In addition, the value candidate of the in point or the out point of the variable set 1 and the value candidate of the variable set 3 include (boundary value analysis) or (equivalent division) as features. “Boundary value analysis” indicates a value candidate based on boundary value analysis. “Equivalent division” indicates a value candidate based on the equivalent division.

  Subsequently, the combination generation unit 17 generates a test case by combining the value candidates generated for the target test item (S180). Specifically, the combination generation unit 17 calculates 1 from each of the other variable sets for each abnormal value candidate of the variable set to which the expression of the target design element of the test item belongs among the variable sets related to the target test item. One normal value candidate is selected to generate test data. However, the boundary value is excluded from the selection target even if it is a normal value. For example, if the target test item is test item 5, the target design element is domain 1, and the expression (5) of domain 1 is the condition of variable set 1, and therefore, three types of variable set 1 in FIG. For each abnormal value candidate, a normal value candidate is selected from each of the variable set 2 and the variable set 3, and a combination of value candidates is generated as test data. Therefore, in this case, three combinations are generated.

  As a result of test case generation, specific values are assigned to all input variables. The combination generation unit 17 generates a test case for each generated combination.

  When steps S130 to S180 are executed for all test items, and further, when steps S120 to S180 are executed for all test objects (business logic), the combination generation unit 17 generates the test case generated in step S180. A test model is generated based on the above and the contents of the generated test model are transmitted to the user terminal 20 as a test case table (S190).

  FIG. 11 is a diagram illustrating an example of a test model. As shown in FIG. 11, the test model includes a test item for each test target. Each test item includes one or more test cases. A test case corresponding to one test item is generated in step S180.

  One test case includes a test case ID, a specific value for each input variable, an expected result, and a test case feature. The test case ID is identification information for each test case. The specific value for each input variable is a combination of value candidates generated in step S180. A plurality of test cases are generated for one test item because there are a plurality of such combinations. The expected result is the result expected by the test. The content of the expected result is generated based on the input constraint or domain definition of the design model. That is, the expected result of the test case related to the test item based on the input constraint is generated based on the “result” of the input constraint. The expected result of the test case regarding the test item based on the domain definition is generated based on the “result” of the domain definition. A test case feature is a feature of a test case. The content of the test case feature is generated based on the feature stored in the value candidate storage unit 111 for the value candidate selected in the test case.

  The test case table indicating the contents of the test model may be generated using, for example, the technique described in Patent Document 2.

  If the structure of the test model described above is strictly defined, for example, as shown in FIG.

  FIG. 12 is a diagram illustrating an example of the definition of the data structure of the test model. In FIG. 12, the data structure of the test model is defined by BNF.

  Next, details of step S140 will be described. FIG. 13 is a flowchart for explaining an example of a processing procedure of input variable classification processing.

  In step S200, the variable classification unit 15 acquires the test data generation conditions extracted by the condition extraction unit 14 in step S130 from the condition extraction unit 14. The subsequent steps S210 to S260 are executed for each test data generation condition. Hereinafter, the test data generation condition that is the processing target is referred to as a “target condition”.

  In step S210, the variable classification unit 15 identifies input variables x1 to xn handled by the target condition (n is the number of input variables handled by the target condition). Subsequently, the variable classification unit 15 determines whether or not the variable set A including all of x1 to xn has been generated (S220). When the variable set A has not been generated (NO in S220), the variable classification unit 15 generates the variable set A and distributes the target condition to the variable set A (S230). Subsequently, the variable classification unit 15 determines whether or not a variable set B that is a subset of the variable set A has been generated (S240). When the variable set B is generated (YES in S240), the variable classification unit 15 distributes the test data generation conditions distributed to the variable set B to the variable set A, and deletes the variable set B (S250). .

  On the other hand, when the variable set A including all of x1 to xn has been generated (YES in S220), the variable classification unit 15 distributes the target condition to the variable set A (S260).

  When steps S210 to S260 are executed for all test data generation conditions, the variable classification unit 15 selects input variables that are not included in any of the already generated variable sets from among the input variables included in the design model. Whether or not there is is determined (S270). When there is one or more corresponding input variables (YES in S270), the variable classification unit 15 generates a variable set for each corresponding input variable (S280). For example, in the present embodiment, the plan name is not included in any test data generation condition. Therefore, the variable set to which the plan name belongs is not generated in step S230, but is generated in step S280.

  The test data generation conditions shown in FIG. 8 will be described with reference to the processing procedure of FIG. First, the expression (1) is set as the target condition. The input variable handled by Equation (1) is specified to be the husband's age (S210). Since a variable set including the husband's age has not been generated (NO in S220), a variable set including the husband's age is generated, and Expression (1) is assigned to the variable set (S230). A variable set that is a subset of the variable set is not generated. Therefore, the expression (2) is set as the target condition. Also for Expressions (2) to (4), each variable set is generated in the same procedure.

  When Expression (5) is set as the target condition, it is specified that the input variables handled by Expression (5) are husband age and wife age (S210). Since a variable set including both husband age and wife age has not been generated (NO in S220), a variable set including both husband age and wife age (hereinafter referred to as “variable set (5)”) is generated. The expression (5) is distributed to the variable set 5 (S230). As a variable set that is a subset of the variable set (5), a variable set generated with respect to the expression (1) (hereinafter referred to as “variable set (1)”) and a variable set generated with respect to the expression (2) ( (Hereinafter referred to as “variable set (2)”) (YES in S240). Therefore, the expressions (1) and (2) are integrated into the variable set (5), and the variable sets (1) and (2) are deleted (S250). As a result, a variable set as shown in FIG. 9 is generated.

  Next, details of step S170 in FIG. 4 will be described. FIG. 14 is a flowchart for explaining an example of a processing procedure of value candidate generation processing using the domain test technique.

  In step S300, the value candidate generator 16 reads one or more test data generation conditions. The test data generation conditions read here (hereinafter referred to as “target conditions”) are test data generation conditions distributed to the target variable set at the time of step S170 in FIG. Subsequently, the value candidate generation unit 16 branches the process based on the data type of the input variable included in the target variable set to which the target condition belongs (S310).

  When all the input variables included in the target variable set are integer types, the value candidate generation unit 16 sets the in point, the out point, the on point, and the off point related to the target condition as value candidates for the input variable. (S320).

  In the present embodiment, when the test item 5 is the target test item and the variable set 1 (husband age, wife age) in FIG. 9 is the target variable set, step S320 is executed.

  On the other hand, when all of the input variables included in the target variable set are character string types, the value candidate generating unit 16 extracts test data generation conditions related to the character string length of the input variables in the target conditions. (S330). Subsequently, the value candidate generator 16 obtains an in point, an out point, an on point, and an off point based on the extracted test data generation conditions (S340). Subsequently, the value candidate generating unit 16 generates a character string that satisfies a target condition other than the character string length as a value candidate for each of the in point, the out point, the on point, and the off point (S350). For example, if there is an input variable (here, string1 and string2) having an input constraint such as “length of string1> = length of string2”, step S350 is executed. Note that if there is no test data generation condition related to the character string length in the condition of interest, step S340 need not be executed.

When all the input variables included in the target variable set are not of any integer type or character string type, such as a real number type, an enumeration type, a date type, or a time type, the value candidate generating unit 16 The boundary value of the condition is converted into an integer type (S360). For example, if it is a real number type, it can be converted to an integer type by multiplying the boundary value by 10 ⁿ (n is the number of digits after the decimal point of the boundary value). In the case of an enumeration type, the enumerated values may be replaced with integers while maintaining the enumeration order relationship. For example, if the listed values are {president, department manager, section manager, section manager, employee}, these may be replaced with {5, 4, 3, 2, 1}. In addition, the boundary value of the target condition may be replaced with the replaced integer. Further, the date type and the time type may be converted into system time such as UNIX (registered trademark) time. For example, a value of “2001-09-09 10:46:40” in Japanese standard time may be converted to a UNIX (registered trademark) time of “1000000000”.

  Subsequently, the value candidate generation unit 16 obtains an in point, an out point, an on point, and an off point based on the focus condition in which the boundary value is converted into the integer type (S370). Subsequently, the value candidate generation unit 16 restores the obtained in point, out point, on point, and off point to the original data type by performing a reverse conversion to step S360. The restored value is a value candidate.

  Subsequent to step S320, S350, or S380, the value candidate generation unit 16 stores the generated value candidate in the value candidate storage unit 111 (S390).

  When the data types of the input variables belonging to the target variable set are different from each other, all the boundary values of the target condition may be converted to the integer type as described in step S360 before step S310. . In this case, after the execution of step S320, the inverse transformation similar to that in step S380 may be performed for each value candidate.

  Next, details of step S320, step S340, and step S370 will be described.

  FIG. 15 is a flowchart for explaining an example of a processing procedure of value candidate generation processing.

  In step S510, the value candidate generation unit 16 obtains an in point of the target condition and sets it as a value candidate, and assigns “normal value (same value division)” as a feature to the value candidate. Subsequently, the value candidate generating unit 16 obtains an out point of the target condition and sets it as a value candidate, and gives “abnormal value (same value division)” as a feature to the value candidate (S520). Subsequently, the value candidate generation unit 16 obtains an on point of the target condition and sets it as a value candidate, and gives “normal value (boundary value analysis)” or “abnormal value (boundary value analysis)” as a feature to the value candidate. (S530). Subsequently, the value candidate generation unit 16 obtains an off point of the target condition and sets it as a value candidate, and assigns “abnormal value (boundary value analysis)” or “normal value (boundary value analysis)” as a feature to the value candidate. (S540).

  Regarding step S530 and step S540, the added feature is not fixed to either “normal value” or “abnormal value”. For the on point and the off point, either the normal value or the abnormal value is determined depending on the condition of interest. Because it changes. Therefore, when the on point or the off point satisfies all the target conditions, the on point or the off point is set to a normal value, and when the on point or the off point does not satisfy any of the target conditions, the on point Alternatively, the off point may be an abnormal value.

  In step S530 and step S540, an on point or an off point may be obtained using a known method. For example, the method described in Patent Document 3 may be used. According to the method described in Patent Literature 3, it is possible to obtain whether the on point is a normal value or an abnormal value, and whether the off point is a normal value or an abnormal value.

  Next, details of step S510 will be described. FIG. 16 is a flowchart for explaining an example of the processing procedure of in point value candidate generation processing.

  In step S600, the value candidate generator 16 reads one or more test data generation conditions (S600). The test data generation conditions read here are the focus conditions at the time of step S510 in FIG.

  Subsequently, the value candidate generation unit 16 generates a constraint equation for obtaining the in point (S610). Specifically, expressions E1 to Em (m is the number of target conditions) constituting the target condition are acquired, and normal values (values that fall within the domain) such as “E1 and E2 and. A constraint equation is generated for obtaining. For example, if the target variable set is the variable set 1 in FIG. 9, a constraint expression “husband age> = 18 and wife age> = 16 and | husband age-wife age | <= 20” is generated.

  Subsequently, the value candidate generation unit 16 obtains a solution of the constraint equation using a known constraint solver, and generates any one solution as an in-point value candidate (S620). When a plurality of input variables belong to the target variable set, in-point value candidates are generated for the plurality of input variables.

  Next, details of step S520 will be described. FIG. 17 is a flowchart for explaining an example of a processing procedure of out point value candidate generation processing.

  In step S700, the value candidate generator 16 reads one or more test data generation conditions (S700). The test data generation conditions read here are the conditions of interest at the time of step S520 in FIG.

  Subsequently, the value candidate generation unit 16 generates a constraint equation for obtaining the out point (S710). Specifically, the expressions E1 to Em (m is the number of the target conditions) constituting the target condition are acquired, and an abnormal value (outside the domain) such as “Not (E1 and E2 and ... and Em)” is obtained. Value) is generated. For example, when the target variable set is the variable set 1 of FIG. 9, a constraint expression “Not (husband age> = 18 and wife age> = 16 and | husband age−wife age | <= 20)” is generated. The

  Subsequently, the value candidate generation unit 16 obtains a solution of the constraint equation using a known constraint solver, and generates any one solution as a value candidate of the out point (S720). When a plurality of input variables belong to the target variable set, out point value candidates are generated for the plurality of input variables.

  Next, details of step S180 in FIG. 4 will be described. FIG. 18 is a flowchart for explaining an example of a processing procedure of test case generation processing.

  In step S400, the combination generation unit 17 reads one or more test data generation conditions. The test data generation conditions read here are all test data generation conditions related to the target test item at the time of step S180 in FIG. For example, when the test item 5 is the target test item, the equations (1) to (5) shown in FIG. 8 are read.

  Subsequently, the combination generation unit 17 acquires all value candidates generated for each variable set related to the target test item from the value candidate storage unit 111 (S410). Subsequently, the combination generation unit 17 acquires the test data generation condition focused on the target test item among the test data generation conditions read in step S400 (S420). The test data generation condition focused on the target test item is a test data generation condition that is a source of generation of the target test item. For example, the test data generation condition focused on in the test item 5 is Expression (5). Hereinafter, the test data generation conditions acquired in step S420 are referred to as “target conditions”.

  Subsequently, the combination generation unit 17 acquires value candidates of a variable set to which one or more input variables handled by the target condition (hereinafter, the one or more input variables are collectively referred to as “variable x”) in step S410. The obtained value candidates are acquired (S430). For example, if the target condition is Expression (5), a value candidate of the variable set 1 (see FIG. 10) is acquired. Subsequently, the combination generation unit 17 selects all value candidates that are abnormal values from the value candidates acquired in step S430 (S440). Whether or not it is an abnormal value can be specified based on whether or not an “abnormal value” is included in the feature of the value candidate.

  Subsequently, the combination generation unit 17 sets the variable set to which each of the remaining input variables other than the variable x (hereinafter referred to as “variables y1 to ym”. M is the number of input variables other than the variable x). A value candidate is acquired from the value candidates acquired in step S410 (S450). Subsequently, the combination generation unit 17 selects any one normal value value candidate for each of the variables y1 to ym from the value candidates acquired in step S450 (S460). Therefore, m value candidates are acquired. However, normal values that are boundary values are excluded as much as possible. This is because, when the test data related to the variable x is generated, if a boundary value is selected for the variables y1 to ym, it becomes difficult to isolate the cause of the bug caused by which input variable when the bug occurs. The boundary value can be specified based on the fact that “value analysis” is included in the feature of the value candidate. In addition, the combination generation unit 17 sets blank input variable value candidates that are not required to be input in the design model (FIG. 6).

  Subsequently, the combination generation unit 17 sets one or more abnormal value candidate values selected in step S440 for the variable x and normal value candidate values selected one by one in step S460 for the variables y1 to ym. All combinations are generated as test cases (S470). At this time, if there are a plurality of variables x, a combination of value candidates is generated for each variable x. For variables y1 to ym, a common value candidate is selected for each combination.

  For example, when the focused test item is the test item 5 and the focused condition is Expression (5), the variable x is the husband age and the wife age. Variables y1 to ym are a plan name and remarks. Therefore, regarding the husband age and the wife age, among the seven value candidates shown in FIG. 10, three different value candidate values are selected. Regarding the plan name, among the three value candidates shown in FIG. 10, any one normal value value candidate is common (fixed) to the three abnormal values related to the husband age and wife age. Selected. Remarks are left blank because they are not essential. That is, the remarks are also common values for the three abnormal values related to the husband age and the wife age. Therefore, there are three combinations of these input variable value candidates. Such three combinations are shown as test cases 5-1 to 5-3 for the test item 5 in FIG. In test cases 5-1 to 5-3 shown in FIG. 11, the value of the plan name and the value of the remarks are common.

  Note that the processing of FIG. 18 is executed for each test item. Each test item is generated for each input constraint or domain definition expression. Therefore, the processing in FIG. 18 is associated with the first variable set to which the input variable (variable x) handled (related to the formula) handled by the formula (target condition) belongs for each formula of the input constraint or domain definition. All conditional expressions associated with the second variable set from each of the second variable sets other than the first variable set for each value candidate that does not satisfy any of the expressions (that is, outliers) This is equivalent to a process of selecting value candidates satisfying (that is, normal values) one by one and generating a combination of value candidates. Furthermore, the value candidates selected from the second group for one expression are common to all the combinations generated for that expression.

  As described above, according to the present embodiment, test data generation conditions are extracted from the design model, and based on the test data generation conditions, it is determined whether or not there is an interdependence between input variables, and based on the determination result. Thus, the input variables are classified into variable sets. When candidate values are generated for each variable set, value candidates are generated by different methods for variable sets to which only one input variable belongs and variable sets to which a plurality of variables having mutual dependency relationships belong. .

  Further, according to the present embodiment, when test data is generated for each test item, the input variable (in FIG. 18) handled by the test data generation condition (the attention condition in FIG. 18) that is the generation source of the test item. With respect to the variable x), abnormal value candidates are comprehensively selected from the value candidates, and regarding the input variables other than the input variable (variables y1 to ym in FIG. 18), one normal among the value candidates. A value is fixedly selected, and a combination of value candidates is generated for each value candidate of the abnormal value.

  As a result, it is possible to comprehensively generate abnormal test data for a plurality of input variables including input variables having a mutual dependency relationship and input variables having no dependency relationship.

  Therefore, for example, when abnormal test data is input to a program, in the test that confirms that the input check operates correctly, it is appropriate to specify which variable is the bug when the bug occurs. Possible abnormal test data can be obtained.

  As mentioned above, although the Example of this invention was explained in full detail, this invention is not limited to such specific embodiment, In the range of the summary of this invention described in the claim, various deformation | transformation・ Change is possible.

DESCRIPTION OF SYMBOLS 10 Test data generation apparatus 11 Design model extraction part 12 Design model analysis part 13 Test item extraction part 14 Condition extraction part 15 Variable classification part 16 Value candidate generation part 17 Combination generation part 20 User terminal 100 Drive apparatus 101 Recording medium 102 Auxiliary storage device 103 memory device 104 CPU
105 Interface Device 111 Value Candidate Storage Unit B Bus

Claims (5)

Based on one or more conditional expressions to be satisfied by a plurality of input variables in the program to be tested, it is determined whether or not there is a dependency relationship between the input variables, and the plurality of input variables are determined based on the presence or absence of the dependency relationship A variable classifying unit that classifies the conditional expression into a group and associates the conditional expression with a group to which an input variable related to the conditional expression belongs;
For each group, based on the conditional expression associated with the group, a value candidate generation unit that generates candidates for the values of input variables belonging to the group;
For each conditional expression, among the candidates generated for the first group to which the input variable related to the conditional expression belongs, for each candidate that does not satisfy any of the conditional expressions associated with the first group, A combination generation unit that selects, from each of the second groups other than the first group, the candidates that satisfy all the conditional expressions associated with the second group one by one, and generates a combination of the candidates. And
The candidate that the combination generation unit selects from the second group for one conditional expression is common to all the combinations generated for the conditional expression.
A test data generation apparatus characterized by the above. The value candidate generation unit generates a candidate for the value of the input variable by a different method for the group including a plurality of input variables and the group including a single input variable.
The test data generation apparatus according to claim 1, wherein: For the group including a plurality of input variables, the value candidate generator generates on-point, out-point, in-point, and off-point based on the conditional expression associated with the group as candidates for the value of the input variable. As sought
When the group including the plurality of input variables is the first group, the combination generation unit selects any one of the out point, the on point, and the off point associated with the group. If the combination is generated for each candidate that does not satisfy the conditional expression and the group including the plurality of input variables is the second group, among the on point, the in point, and the off point, Selecting candidates that satisfy all the conditional expressions associated with the group;
The test data generation apparatus according to claim 1 or 2, wherein Computer
Based on one or more conditional expressions to be satisfied by a plurality of input variables in the program to be tested, it is determined whether or not there is a dependency relationship between the input variables, and the plurality of input variables are determined based on the presence or absence of the dependency relationship A variable classification procedure for classifying the conditional expression into a group and associating each conditional expression with a group to which an input variable related to the conditional expression belongs;
For each group, based on the conditional expression associated with the group, a value candidate generation procedure for generating candidates for values of input variables belonging to the group;
For each conditional expression, among the candidates generated for the first group to which the input variable related to the conditional expression belongs, for each candidate that does not satisfy any of the conditional expressions associated with the first group, A combination generation procedure for selecting the candidates satisfying all the conditional expressions associated with the second group one by one from each of the second groups other than the first group, and generating the combination of the candidates. And run
The candidate selected from the second group for one conditional expression in the combination generation procedure is common to all the combinations generated for the conditional expression.
A test data generation method characterized by the above. On the computer,
Based on one or more conditional expressions to be satisfied by a plurality of input variables in the program to be tested, it is determined whether or not there is a dependency relationship between the input variables, and the plurality of input variables are determined based on the presence or absence of the dependency relationship A variable classification procedure for classifying the conditional expression into a group and associating each conditional expression with a group to which an input variable related to the conditional expression belongs;
For each group, based on the conditional expression associated with the group, a value candidate generation procedure for generating candidates for values of input variables belonging to the group;
For each conditional expression, among the candidates generated for the first group to which the input variable related to the conditional expression belongs, for each candidate that does not satisfy any of the conditional expressions associated with the first group, A combination generation procedure for selecting the candidates satisfying all the conditional expressions associated with the second group one by one from each of the second groups other than the first group, and generating the combination of the candidates. And let the computer run
The candidate selected from the second group for one conditional expression in the combination generation procedure is common to all the combinations generated for the conditional expression.
A test data generation program characterized by that.

Images