Equivalence class partioning and boundary value analysis

Data Driven Testing uses equivalence class decomposition and boundary value analysis to efficiently reduce test cases without compromising coverage. In this article, we show how these methods are combined to enable more precise tests with optimized effort.

Basics of the equivalence class method

The equivalence class method is a test case derivation technique that aims to divide the input data of a system into groups or classes in which the values can be considered equivalent. Equivalent means that each value within a class should behave identically in the test case behavior, whereby it is sufficient to test with only one representative of each class.

Example of an equivalence class division

Assume a form accepts numbers between 1 and 100 as valid input:

Input values	Equivalence class	Validity
-10, 0	invalid	No
1 – 100	valid	Yes
101, 150	invalid	Nein

Combination with the boundary value analysis

Boundary value analysis is a complementary technique to the equivalence class method that focuses on testing the edges of the equivalence classes. As errors often occur at the limits of a valid value range, these values are particularly critical for tests.

Example of the boundary value analysis

For the above numerical range (1 – 100), the relevant limit values could be defined as follows:

Input values	Classification	Expected result
0	Lower boundary -1	Defect
1	Lower boundary	Success
100	Upper boundary	Success
101	Upper boundary +1	Defect

The combination of both methods ensures a more precise test coverage by testing representative values from each equivalence class as well as critical boundary values.

Application in Data Driven Testing

Data Driven Testing is based on the idea of executing a large number of test cases with different input values. The equivalence class method helps to optimize the amount of test data.

Advantages of combining DDT with equivalence classes and boundary value analysis

1. Reduction of redundant tests: Instead of testing all possible inputs, only representative values from each equivalence class and critical boundary are selected.
2. Reduction of redundant tests: Instead of testing all possible inputs, only representative values from each equivalence class and critical boundary are selected.
3. Higher test coverage with fewer test cases: Critical test cases are covered without performing unnecessary tests.
4. Higher test coverage with fewer test cases: Critical test cases are covered without performing unnecessary tests.
5. Higher test coverage with fewer test cases: Critical test cases are covered without performing unnecessary tests.

Visualization of equivalence classes and boundray value analysis

A diagram can illustrate the classification of the equivalence classes and the relevant boundary values:

• X-axis: Input values
• Y-axis: Validity

Test data matrix for data-driven testing with limit value analysis

Test case	Input values	Expected result
TC1	-10	Defect
TC2	0	Defect
TC3	1	Success
TC4	50	Success
TC5	100	Success
TC6	101	Defect

Example 1: Age check for the registration of an online platform

Suppose an online platform only allows users between the ages of 18 and 65 to register. The software validates the age based on the number entered by the user.

Equivalence class analysis

Input values	Equivalence class	Validity
0 – 17	Invalid	No
18 – 65	Valid	Yes
66 und höher	Invalid	No

Boundary value analysis

Here we specifically test the values at the boundary of the equivalence classes:

Test case	Input values	Expected result
TC1	17	Defect (too young)
TC2	18	Success (currently valid)
TC3	30	Success (currently valid)
TC4	65	Success (just valid)
TC5	66	Defect (too old)

Explanation of the test strategy

• By combining the equivalence class method and boundary value analysis, we minimize the test cases while still ensuring maximum defect coverage.
• Instead of testing all ages between 18 and 65, we choose only representative values.
• The thresholds 17, 18, 65, 66 are critical as defects in the implementation often occur here.

This example shows how these simple methods help to efficiently test the quality of an application in practice.

The example above shows the simplest partioning into equivalence classes: a class with valid and a class with invalid representatives. The second example combines several data types with several equivalence classes.

Example 2: Flight ticket prices based on age and booking class

A flight booking system calculates the ticket price depending on the age of the passenger and the booking class. The prices and booking conditions are as follows:

• Children (0-11 years) receive a 50% discount, but may only travel with an adult.
• Teenagers (12-17 years) pay the full price.
• Adults (18-64 years) pay the full price.
• Seniors (65+ years) receive a 20% discount.
• Normal prices apply in Economy class.
• In Business Class there is a surcharge of 50% on the base fare.
• In First Class there is a surcharge of 100% on the base fare.

Equivalence class analysis

Here we divide the possible inputs into equivalence classes for two data types:

Age of the passenger	Equivalence class	Permitted booking	Discount/surcharge
0 – 11 Jahre	Kids	Only with adult	-50%
12 – 17 Jahre	Teenagers	Yes	No discount
18 – 64 Jahre	Erwachsene	Yes	No discount
65+ Jahre	Senioren	Yes	-20%

Booking class	Equivalent class	Price rule
Economy	Standard	Normal price
Business	Premium	+50%
First Class	Luxus	+100%

Boundary value analysis

Here we specifically test the boundary values between the equivalence classes:

Test Case	Input Values	Expected result
TC1	0	Defect (child without adult)
TC2	1	Defect (child without adult)
TC3	11	Defect (child without adult)
TC4	12	Success (transition child → teenager)
TC5	17	Success (last teenage stage)
TC6	18	Success (transition teenager → adult)
TC7	64	Success (last adult stage)
TC8	65	Success (transition adult → senior with 20% discount)
TC9	80	Success (Senior with 20% discount)
TC10	30	Success (50% surcharge)
TC11	40	Success (100% surcharge)

Explanation of the test strategy

• Several equivalence classes for age and booking class.
• Boundary value analysis checks whether the system processes the transition between age groups correctly.
• Special case: Children must travel with an adult, therefore TC1-TC3 are invalid.
• Discount and surcharge calculations are checked.

Conclusion

Combining the equivalence class method with boundary analysis is a valuable technique in Data Driven Testing to select test cases and increase test efficiency. While equivalence classes sensibly reduce the amount of test data, boundary value analysis ensures that critical edge cases are not overlooked. This systematic selection of test data can ensure the quality of the software and minimize the testing effort at the same time.

Search