CN107861873A - Priorities of test cases method of adjustment based on the adjustment of two attribute hierarchies - Google Patents

Abstract

The public priorities of test cases method of adjustment based on the adjustment of two attribute hierarchies of the present invention, comprises the following steps：Step 1, the division and extraction to test use cases；Step 2, the decoupling of test use cases division result；Step 3, priority dynamic adjusts.Priorities of test cases method of adjustment of the present invention establishes incidence relation by the use-case for concentrating test case and two attributes, by decoupling and dividing adjustment grade, enter Mobile state adjustment priority to use-case in measuring executing process, it the method increase the accuracy of test cases selection to be adjusted, the run time that mistake is found in test can be shortened, testing efficiency is improved, there is good practical value.

Description

A test case priority adjustment method based on two-attribute hierarchical adjustment

technical field

The invention belongs to the technical field of computer software testing methods, and in particular relates to a test case priority adjustment method based on two-attribute hierarchical adjustment.

Background technique

In recent years, people's demand for computers is increasing. In order to meet people's needs, software developers have developed more life software, which makes the number of mobile applications grow rapidly and the types of mobile applications become more diverse. As the application software has more and more functions, the business logic inside and between the software modules is also becoming more and more complex, which greatly increases the probability of software defects. Therefore, it is very necessary to test a part of the software after completing its functions. However, due to the short time of software development in actual testing, the time left for the testing process is very limited. Executing test cases in accordance with the original sequence when testing a software will virtually prolong the time for finding errors and reduce testing efficiency.

Each test case is not isolated, and there may be some relationship between them. According to this feature, the use cases to be executed in the test case set can be linked with the use cases where defects are found, and the execution order of these use cases can be adjusted during execution. , the purpose is to find the same or similar defects as early as possible. At the same time, there are many attributes of test cases, and which attribute to choose to divide the test case set determines the closeness of the test cases in the same "domain". The adjustment range of the use case will affect the time to find the defect. Only by finding the appropriate division attribute and adjustment scheme can the time to find the error be effectively shortened.

Contents of the invention

The purpose of the present invention is to provide a test case priority adjustment method based on two-attribute hierarchical adjustment, which solves the problem of coupling between test cases and their attributes in existing software, selection of test cases to be adjusted is not targeted, and adjustment efforts are single question.

The technical solution adopted in the present invention is a test case priority adjustment method based on two-attribute hierarchical adjustment, comprising the following steps:

Step 1, division and extraction of test case sets

Divide the test case set according to the design meaning, and then extract the test case set with a common function type;

Step 2, decoupling of test case set division results

Drawing on the idea of the scenario method, the "result scenario" is added after the end of the operation event flow: Since the designer not only gives the expected result of each test case's execution success, but also predicts the possible routine scenarios of its execution failure, the introduction of "can Predictable Failure Result Scenarios” provides its unique “predictable failure result scenarios” for each test case corresponding to the design meaning; by comparing the actual failure scenarios with the “predictable failure result scenarios”, improve the accuracy of test case execution failures. The accuracy of the design significance judgment;

Step 3, priority dynamic adjustment

First, classify unexecuted use cases into first-level promotion use cases, second-level promotion use cases, and temporary non-adjustment use cases; then perform initialization sorting according to the historical execution status of each test case; finally, adjust the priority according to the initialization sorting results.

The present invention is also characterized in that,

The specific operation of step 1 is:

Step 1.1, the division of test case sets should follow the following principles: multiple test cases in a test case set can come from the same design meaning, and a test case in a test case set can also contain or cover multiple design meanings;

In step 1.2, the following principles should be followed for extracting the test case sets with common function types: when the test cases belong to different functional modules but contain the same operation function, multiple test cases can belong to the same function type, and the remaining one test case A use case can only cover one functional type.

The specific operation of step 2 is:

Let any design meaning be d, the "predictable failure result scenario" be fau(d), and the set of "predictable failure result scenarios" be Fau(d), then Fau(d)={fau(d) ₁ ,fau( d) ₂ ,...,fau(d) _m }, where m≥1;

The elements in the failure scenario collection are obtained by enumeration, and the number of elements can be one or more; if a test case fails to execute, and this test case covers multiple design meanings, the actual The failure scenario is matched one by one with the elements in the failure scenario set of each design meaning contained in it. If the actual failure scenario matches one of the elements successfully, the set to which this element belongs is the "predictable failure" of the target design meaning. Result scenarios" set Fau(d), and then determine the target design meaning.

The specific operation process of step 3 is:

Step 3.1, unexecuted use case classification: If a test case fails to execute during the test, the test case that belongs to the same functional type and contains the same design meaning as the failed test case is classified as a first-level promotion use case; Test cases that belong to the same function type or contain the same design meaning are classified as second-level promotion cases; test cases other than the above two are classified as temporary non-adjustment cases;

Step 3.2, according to the historical execution status of the test cases and the classification order of the functional modules, initialize and sort each test case in step 3.1;

Step 3.3, according to the initialization sorting in step 3.2, start to execute the test cases with the highest priority. If a test case fails to execute during the test, that is, the execution result is different from the expected result, first increase the priority of the test case in step 3.1. After upgrading to the current execution use case, if the test case only covers one design meaning, then directly calculate the second-level promotion use case in step 3.1 according to formula (1), and adjust the degree of improvement; if the test case covers multiple design meanings , then first determine the meaning of the target design by using step 2, and then adjust the calculation of the improvement rate for the second-level improvement use case in step 3.1 according to formula (1);

Δrange=c(caseNum-exeNum-lv1Num) (1)

Among them, Δrange is the number of grades that need to be upgraded; caseNum is the total number of use cases contained in the test case set; exeNum is the number of executed use cases; lv1Num represents the number of use cases that have been determined as a first-level upgrade; c is the grade upgrade coefficient, Used to control the size of the promotion level, and c∈(0,1).

The historical execution situation in step 3.2 is specifically the error detection rate or error level of the test case.

The beneficial effects of the present invention are: a test case priority adjustment method based on two-attribute hierarchical adjustment of the present invention establishes an association relationship between the use cases in the test case set and the two attributes, and after decoupling and division and adjustment levels, in the test execution process In this method, the priority of use cases is dynamically adjusted. This method improves the accuracy of the selection of test cases to be adjusted, can shorten the running time of errors found in the test, and improves the test efficiency. It has very good practical value.

Description of drawings

Fig. 1 is the flow chart of the test case priority adjustment method based on two attribute classification adjustments of the present invention;

Fig. 2 is the flowchart of "result scene" in the test case priority adjustment method of the present invention;

Fig. 3 is the relationship diagram of the adjustment of test case priority of the present invention;

Fig. 4 is a comparison diagram of the adjustment method of the present invention and other priority algorithms;

Fig. 5 is a comparison chart of average defect discovery rate (APFD) values between the adjustment method of the present invention and other priority algorithms.

Detailed ways

The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments.

A kind of test case priority adjustment method based on two-attribute hierarchical adjustment of the present invention, as shown in Figure 1, comprises the following steps:

Step 1, division and extraction of test case sets

In step 1.1, the division of test case sets should follow the following principles:

(1) Multiple test cases in the test case set can come from the same design meaning, such as the commonly used test method - equivalence class division method, the test cases in the effective equivalence class reflect whether the test program can achieve the predetermined function Purpose, the test cases in the invalid equivalence class all reflect the purpose of whether the test program can avoid the exception caused by invalid input; (2) An execution action may produce multiple operation results, and these operation results are used to judge whether a function is perfect Therefore, one test case in the test case set can also contain or cover multiple design meanings. Based on the above principles, as shown in Table 1, the following test case-design significance correlation matrix example is given:

Table 1 Use Case-Design Meaning Correlation Matrix

In step 1.2, the extraction of test case sets with common functional types should follow the following principles:

When the test cases belong to different functional modules but contain the same operation function, multiple test cases can belong to the same function type, and the remaining test cases can only cover one function type.

The function type that is contained in at least two modules is called the common function type fc, and a common function type set FC={fc ₁ ,fc ₂ ,...,fc _k } is extracted, where k≥0, the common function type The set FC contains all common function types of the cases to be tested. Assign the use cases in the test case set to each element in the FC to the greatest extent, and the remaining unassigned test cases, that is, those containing the unique functions of the module, no longer participate in the judgment and adjustment of the function type. Table 2 shows an example of the test case-function type correlation matrix:

Table 2 Use Case-Functional Type Correlation Matrix

Step 2, decoupling of test case set division results

As shown in Figure 2, drawing on the idea of the scenario method, the "result scenario" is added after the end of the operation event flow: because the designer not only gives the expected result of the successful execution of each test case, but also predicts the possible routine of its execution failure. Therefore, "predictable failure result scenarios" are introduced to provide its unique "predictable failure result scenarios" for each test case corresponding to the design meaning; by comparing the actual failure scenarios with the "predictable failure result scenarios", improve The accuracy of judgments about the design significance of the test case execution failure;

Let any design meaning be d, the "predictable failure result scenario" be fau(d), and the set of "predictable failure result scenarios" be Fau(d), then Fau(d)={fau(d) ₁ ,fau( d) ₂ ,...,fau(d) _m }, where m≥1;

The elements in the failure scenario collection are obtained by enumeration, and the number of elements can be one or more; if a test case fails to execute, and this test case covers multiple design meanings, the actual The failure scenario is matched one by one with the elements in the failure scenario set of each design meaning contained in it. If the actual failure scenario matches one of the elements successfully, the set to which this element belongs is the "predictable failure" of the target design meaning. Result scenarios" set Fau(d), and then determine the target design meaning.

Step 3, priority dynamic adjustment

Step 3.1, unexecuted use case classification: If a test case fails to execute during the test, the test case that belongs to the same functional type and contains the same design meaning as the failed test case is classified as a first-level promotion use case; Test cases that belong to the same function type or contain the same design meaning are classified as second-level promotion cases; test cases other than the above two types are classified as temporary non-adjustment cases, as shown in Figure 3, the test case promotion level is represented by a set relationship relation;

Step 3.2, determine the initial priority of each test case according to the historical execution of the test case, such as error detection rate, error level, etc., or classify it as a function module first, and the overall order of the function modules shall prevail. Step 3.1 Each test case in the initialization sequence;

Step 3.3, according to the initialization sorting in step 3.2, start to execute the test cases with the highest priority. If a test case fails to execute during the test, that is, the execution result is different from the expected result, first increase the priority of the test case in step 3.1. After upgrading to the current execution use case, if the test case only covers one design meaning, then directly calculate the second-level promotion use case in step 3.1 according to formula (1), and adjust the degree of improvement; if the test case covers multiple design meanings , then first determine the meaning of the target design by using step 2, and then adjust the calculation of the improvement rate for the second-level improvement use case in step 3.1 according to formula (1);

Δrange=c(caseNum-exeNum-lv1Num) (1)

Among them, Δrange is the number of grades that need to be upgraded; caseNum is the total number of use cases contained in the test case set; exeNum is the number of executed use cases; lv1Num represents the number of use cases that have been determined as a first-level upgrade; c is the grade upgrade coefficient, Used to control the size of the promotion level, and c∈(0,1). The setting of the c value follows the following principles: the value of c is determined by the tester depending on the specific software to be tested. If the functional modules of the software to be tested are closely related or have similar functions, for example, the business logic of each management module in the management system has very If the similarity is large, then c can take a larger value; if there is no outstanding similarity between the functional modules of the software to be tested, for example, there is no obvious functional feature or it contains a variety of functions, then c can take a smaller value .

As shown in Figure 4, the present invention is based on the two-attribute hierarchical adjustment algorithm (TA&HA) and the whole process unsorted (Unsorted) and functional domain division algorithm (FUDD) in the test, the comparison diagram of the relationship between the test case execution rate and the error detection rate , it can be seen from Figure 4 that when testing the software with the same number of defects, the TA&HA algorithm can find more defects than the other two algorithms in the case of executing the same number of test cases, and the time to find errors is earlier.

As shown in Figure 5, the present invention is based on the APFD value comparison chart of the two-attribute hierarchical adjustment algorithm (TA&HA) and the whole process unsorted (Unsorted) and functional domain division algorithm (FUDD). The algorithm has a high APDF value, has higher adjustment efficiency, and can find software defects faster.

The priority adjustment method of the present invention locates and selects use cases to be adjusted from the two attributes of design meaning and function type, and divides the adjustment range of test cases into two levels, which are graded and adjusted according to actual conditions during the execution process, and have a more Fine selection and adjustment strategies can make software defects exposed in a short period of time, which can effectively reduce testing costs and improve the execution efficiency of test cases.

Claims (5)

1. the priorities of test cases method of adjustment based on the adjustment of two attribute hierarchies, it is characterised in that comprise the following steps：

Step 1, the division and extraction to test use cases

According to design sense partition testing set of uses case, the test use cases for possessing shared function type are then extracted；

Step 2, the decoupling of test use cases division result

The thought of scene method is used for reference, " result scene " is supplemented after Action Events stream terminates：Due to designer not only provide it is each The expected result that test case runs succeeded, and pre- cicada its conventional scenario for unsuccessfully being likely to occur of execution, therefore introducing " can Predict failure result scene ", providing its for test case corresponding to each design sense exclusive " can predict failure result field Scape "；By contrasting actual failure scene and " failure result scene can be predicted ", improve and set to causing test case to perform failure Count the accuracy that meaning judges；

Step 3, priority dynamic adjusts

Use-case classifying and dividing will be not carried out first to lift use-case, two-stage hoisting use-case for one-level and use-case wouldn't be adjusted；Then Initialization sequence is carried out according to the history implementation status of each test case；Finally priority tune is carried out according to initialization ranking results It is whole.

2. the priorities of test cases method of adjustment according to claim 1 based on the adjustment of two attribute hierarchies, its feature exist In the specific gymnastics conduct of the step 1：

Step 1.1, partition testing set of uses case should follow following principle：Test case concentrates multiple test cases can be from same Individual design sense, test case concentrate a test case also can include or cover multiple design senses；

Step 1.2, extract and possess the test use cases of shared function type and should follow and ask following principle：When belonging to difference in functionality During the situation of module but test case comprising same operation function, multiple test cases can belong to same function type, its A test case of remaininging can only cover One function type.

3. the priorities of test cases method of adjustment according to claim 1 based on the adjustment of two attribute hierarchies, its feature exist In step 2 concrete operations are：

If any design sense is d, " can predict failure result scene " is fau (d), " can predict failure result scene " collection is combined into Fau (d), then Fau (d)={ fau (d)₁,fau(d)₂,...,fau(d)_m, wherein m >=1；

Element in failure scene set is obtained by enumerating mode, and its element number can be one or more；If a certain survey Example on probation performs failure, and this test case covers multiple design senses, then by the actual failure scene of the test case, The element in failure scene set with its each design sense included is matched one by one, if actual failure scene with wherein A certain Match of elemental composition success, then " failure result scene can be predicted " collection of the set belonging to this element, as target design meaning Fau (d) is closed, and then determines target design meaning.

4. the priorities of test cases method of adjustment according to claim 1 based on the adjustment of two attribute hierarchies, its feature exist In the specific operation process of the step 3 is：

Step 3.1, it is not carried out use-case classifying and dividing：If in test process, a certain test case performs failure, then with the failure Test case had both belonged to same function type, and the test case comprising same design meaning is classified as one-level lifting use-case；Will category In same function type, or test case comprising same design meaning is classified as two-stage hoisting use-case；Outside both the above Test case is classified as that use-case wouldn't be adjusted；

Step 3.2, according to the history implementation status of test case and functional module classification order, to each test in step 3.1 Use-case carries out initialization sequence；

Step 3.3, sorted according to initialization in step 3.2, the implementation of test cases since highest priority, if testing Certain use-case performs failure in journey, i.e. implementing result is different from expected results, then one-level in step 3.1 first is lifted into the preferential of use-case Level lifting is to after currently performing use-case, if the test case only covers a design sense, directly by step 3.1 two Level lifting use-case is calculated according to formula (1), and lifting amplitude is adjusted；If the test case covers multiple design senses, Then first pass through and determine target design meaning using step 2, then by two-stage hoisting use-case in step 3.1 according to formula (1), calculating carries Increasing degree degree is adjusted；

Δ range=c (caseNum-exeNum-lv1Num) (1)

Wherein, Δ range is the number of degrees for needing to be lifted；CaseNum is that test case concentrates the total use-case number included； ExeNum is the use-case number of executed；Lv1Num represents to have been determined as the number of one-level lifting use-case；C is grade lifting system Number, for controlling the size of lifting grade, and c ∈ (0,1).

5. the priorities of test cases method of adjustment according to claim 4 based on the adjustment of two attribute hierarchies, its feature exist In history implementation status is specially the error detection rate or errorlevel of test case in the step 3.2.