Line Coverage vs Branch Coverage

When discussing code coverage, two of the most common metrics are line coverage and branch coverage. While both are important, they measure different aspects of how thoroughly your code is tested.

Line Coverage

Line coverage (also called statement coverage) is the most basic and commonly used code coverage metric. It measures the percentage of executable lines of code that are executed when your tests run.

Line Coverage = (Number of Lines Executed / Total Number of Executable Lines) × 100%

Advantages:

• Simple to understand and calculate
• Provides a good high-level overview of test coverage
• Easy to visualize in coverage reports

Limitations:

• Doesn't account for decision points (if/else, switch statements, etc.)
• Can show high coverage even when critical paths aren't tested
• May give a false sense of security when complex logic is involved

Branch Coverage

Branch coverage measures whether each possible branch from a decision point has been executed. For example, in an if-else statement, branch coverage would require tests that execute both the "if" and the "else" paths.

Branch Coverage = (Number of Branches Executed / Total Number of Branches) × 100%

Advantages:

• More thorough than line coverage
• Better at finding untested decision paths
• Helps ensure all logical paths through the code are tested
• More effective for complex conditional logic

Limitations:

• More complex to understand and implement
• Typically requires more tests to achieve high coverage
• Doesn't guarantee all combinations of conditions are tested

Line vs Branch Coverage: An Example

Consider this PHP code snippet:

function getDiscount($amount, $isVIP) {
    $discount = 0;  // Line 1
    
    if ($amount > 100 && $isVIP) {  // Line 2
        $discount = 0.2;  // Line 3
    } else if ($amount > 100) {  // Line 4
        $discount = 0.1;  // Line 5
    }
    
    return $discount;  // Line 6
}

A test that calls getDiscount(150, true) would execute lines 1, 2, 3, and 6, giving a line coverage of 4/6 = 66.7%. However, it would only cover 1 of 3 branches (the first condition being true), giving a branch coverage of only 33.3%.

To achieve 100% branch coverage, you would need additional tests like getDiscount(150, false) and getDiscount(50, false).

Which Should You Use?

Ideally, you should track both line and branch coverage:

• Line coverage gives you a broad view of which parts of your code are executed
• Branch coverage ensures your tests examine different decision paths

Many coverage tools report both metrics, and some also provide more advanced measurements like path coverage or condition coverage for even more thorough testing.

OtterWise helps you track these coverage metrics over time, ensuring your test suite remains comprehensive as your codebase evolves.