Wsrtjtyk

So I am in charge of creating a project that has to do with calculations of pay. One aspect of the project is to analyze a proper input such as 4:00 PM, and other aspects including calculating the pay for the hours put in, and the type of job etc.

my question more so has to do with the best practices for designing the classes around this.

Should I have one class that analyzes the input string, and only does that? and one class for the calculator to display the proper output or should it all be in one class?

both ways are fine for me to do, but what is considered professional?

is it best practice to split classes based on their unique functionality even if you dedicate a class to simply one method?

At the boundary of your application you'll be accepting requests through a user interface or a system interface. You should treat anything originating from outside your application as untrusted and potentially wrong. For example, if you receive a HTTP request there is no guarantee that it is valid and contains the fields you expect. If you read form a file, it might be incorrectly formatted.

There should be a layer at the boundary of your application which takes input (which is just a bunch of bytes in the end) and turns it into a representation as Java objects of the suitable type (e.g. Boolean, LocalDate). If everything is a String, you are probably doing it wrong.. If this layer is unable to do this, it should send back an error.

Once you have expressed the request as a correctly typed Java objects, your business logic should process the request. This makes it possible to use the same logic when data is provided through a different interface, separates plumbing (parsing) from business logic (calculations). It allows the business logic to be more easily unit tested.

When you output a response back to the user (or system), you should convert from your nicely structured Java objects back to the output representation at the last moment.

I suggest you take a look at the javax.validation package and the Bean Validation JSR's 1.0 and 2.0

Using this approach you can create Java classes to represent your data and annotate them with the required validations. Triggering the validation to happen depends a little bit on the context.

In a Spring Boot application putting @Valid on the received controller parameter does the trick. See also this cheat sheet:

  import javax.validation.Valid;

  import com.company.app.model.Article;



  @Controller

  public class ArticleController {



  ...



  @RequestMapping(value="/postArticle", method=RequestMethod.POST)

    public @ResponseBody String postArticle(@Valid Article article, BindingResult result, HttpServletResponse response){



        if(result.hasErrors()){

            String errorMessage = "";

            response.setStatus(HttpServletResponse.SC_BAD_REQUEST);

            List<ObjectError> errors = result.getAllErrors();



            for( ObjectError e : errors){

                errorMessage+= "ERROR: " + e.getDefaultMessage();

            }

            return errorMessage;

        }



        else{

               return "Validation Successful";  

        }

      }

  }

In a standalone application it could be done like this:

public class BeanValidationExample {

  public static void main (String args) {

    Configuration<?> config = Validation.byDefaultProvider()

                                        .configure();

    ValidatorFactory factory = config.buildValidatorFactory();

    Validator validator = factory.getValidator();

    factory.close();



    Person person = new Person();

    person.setDateOfBirth(new Date(System.currentTimeMillis() + 10000));



    Set<ConstraintViolation<Person>> violations = validator.validate(person);

       violations.forEach(v -> System.out.println(v.getPropertyPath() +

                                                      "- " + v.getMessage()));

  }

}