adam bien's blog

Sophisticated business logic requires a high amount of high quality data for test purposes. Tabular data is easier maintainable in Excel then directly in Java code.

With Java 8 it is easy to provide a function which converts an Excel Row (actually a HSSF Row) into whatever POJO you like:

 Function pojoMapper() {
        return (row) -> {
            Iterator cells = row.cellIterator();
            return new Input(
                    asLong(cells.next()),
                    asLong(cells.next()),
                    asLong(cells.next()));
        };
    }

A few additional lines of code convert a Stream<Row> into a Stream<POJO>:

public static <T> Stream<T> load(Function<Row, T> mapper,...) {
        int skipCount = 0;
        if (hasHeader) {
            skipCount = 1;
        }
        String fileName = //...
        try (InputStream inp = new BufferedInputStream(new FileInputStream(fileName));) {
            try (XSSFWorkbook wb = new XSSFWorkbook(inp)) {
                XSSFSheet sheet = wb.getSheetAt(tab);
                Stream<Row> stream = StreamSupport.stream(sheet.spliterator(), false);
                return stream.skip(skipCount).map(mapper);

            }
        } catch (IOException ex) {
            throw new IllegalStateException("Problems processing file: " + fileName, ex);
        }
    }

Parameterized JUnit tests can be used to define the locations of excel files and load the sheets representing different use cases, test suites or regression tests.

<dependency>
	<groupId>com.airhacks</groupId>
	<artifactId>sheetfit</artifactId>
	<version>0.0.1</version>
</dependency>

Example is implemented as system test: https://github.com/AdamBien/sheetfit/tree/master/sheetfit-st.

Also checkout javaeetesting.com or come to MUC: http://workshops.adam-bien.com/about-testing.htm