OpenEDG PCAD: Programming Practices: Object-Oriented Design and Exception Handling

What this domain tests

Programming practices: object-oriented design and exception handling in data code. The exam checks whether your classes have clear responsibilities and whether failures are handled deliberately rather than swallowed.

Core rules

• A class bundles state and behavior: __init__ sets attributes, and methods operate on them. Prefer composition, such as a loader object used by a pipeline, over deep inheritance.
• Encapsulation: keep implementation attributes private by convention with a leading underscore and expose methods.
• Exceptions: catch only what you can handle, name specific types such as except ValueError:, and use else for code that runs when no exception occurred and finally for cleanup.
• Raise exceptions with context: raise ValueError(f'missing column: {col}').
• Use context managers such as with open(...) and with sqlite3.connect(...) to guarantee release.

Worked example

A DataLoader class wraps a file path. Its load() method opens the file with with, parses JSON, and raises a custom DataFormatError if the root is not a list. The caller catches DataFormatError, logs the path, and skips that file; it does not catch Exception, so unexpected bugs still surface. Reasoning: the class owns one responsibility, loading; the custom exception communicates a domain-specific failure; and the narrow catch prevents masking programming errors.

Traps that cost marks

• except Exception: pass, which hides real bugs.
• Catching an exception and continuing with a partially built object.
• Using a mutable default argument such as def f(items=[]) and sharing state across calls.
• Forgetting to close files or database connections on the error path.
• Building a deep inheritance hierarchy when a small function or composition would be clearer.