Impara Data vs Non-Data Descriptors

Scorri per mostrare il menu

Not all descriptors behave the same way in the attribute lookup chain. Python distinguishes between data descriptors and non-data descriptors based on which methods they define – and this distinction determines whether a descriptor can be shadowed by an instance attribute.

The Distinction

A data descriptor defines __set__ or __delete__ (or both) in addition to __get__. It takes priority over the instance's __dict__ in the lookup chain.

A non-data descriptor defines only __get__. It can be shadowed by an instance attribute with the same name.


              123456789101112131415161718192021222324252627282930
            
# Data descriptor – cannot be shadowed by instance __dict__
class DataDescriptor:
    def __get__(self, obj, objtype=None):
        if obj is None:
            return self
        return obj.__dict__.get("_value", "no value")

    def __set__(self, obj, value):
        obj.__dict__["_value"] = value

# Non-data descriptor – can be shadowed
class NonDataDescriptor:
    def __get__(self, obj, objtype=None):
        if obj is None:
            return self
        return "from descriptor"

class Example:
    data = DataDescriptor()
    non_data = NonDataDescriptor()

example = Example()

# Trying to shadow the data descriptor
example.__dict__["data"] = "shadowed attempt"
print(example.data)      # Still calls __get__ – data descriptor wins

# Shadowing the non-data descriptor
example.__dict__["non_data"] = "instance value"
print(example.non_data)  # "instance value" – instance __dict__ wins

Why the Distinction Matters

Data descriptors are used when you want to enforce behavior on every read and write – validation, type checking, computed properties. The instance cannot bypass the descriptor by setting an attribute directly.

Non-data descriptors are used for behavior that should be overridable per instance – functions are the primary example (covered in Chapter 4).

The Priority Table

A Practical Example

A data descriptor for validated string attributes that cannot be bypassed:


              12345678910111213141516171819202122232425262728
            
# Data descriptor enforcing non-empty strings
class NonEmptyString:
    def __set_name__(self, owner, name):
        self._name = name

    def __get__(self, obj, objtype=None):
        if obj is None:
            return self
        return obj.__dict__.get(self._name, "")

    def __set__(self, obj, value):
        if not isinstance(value, str) or not value.strip():
            raise ValueError(f"{self._name} must be a non-empty string, got {value!r}")
        obj.__dict__[self._name] = value.strip()

class Employee:
    name = NonEmptyString()
    department = NonEmptyString()

    def __init__(self, name, department):
        self.name = name
        self.department = department

employee = Employee("Alice", "Engineering")
print(employee.name)        # "Alice"

# Trying to bypass via __dict__ has no effect – __set__ is always called
employee.name = ""          # Raises ValueError

Read-Only Data Descriptors

A descriptor with __get__ and __delete__ but no __set__ is also a data descriptor – and attempting to set it raises AttributeError:


              123456789101112131415161718192021222324
            
# Read-only data descriptor using __set__ to raise AttributeError
class ReadOnly:
    def __set_name__(self, owner, name):
        self._name = "_" + name

    def __get__(self, obj, objtype=None):
        if obj is None:
            return self
        return obj.__dict__.get(self._name)

    def __set__(self, obj, value):
        if self._name in obj.__dict__:
            raise AttributeError(f"{self._name[1:]} is read-only")
        obj.__dict__[self._name] = value  # Allowing the initial set from __init__

class Transaction:
    transaction_id = ReadOnly()

    def __init__(self, transaction_id):
        self.transaction_id = transaction_id  # Initial set allowed

transaction = Transaction("TX-001")
print(transaction.transaction_id)  # "TX-001"
transaction.transaction_id = "TX-002"  # Raises AttributeError

Tutto è chiaro?

Grazie per i tuoi commenti!

Sezione 1. Capitolo 3

Chieda ad AI

Chieda pure quello che desidera o provi una delle domande suggerite per iniziare la nostra conversazione

Sezione 1. Capitolo 3