Swipe um das Menü anzuzeigen

Every Python function is a non-data descriptor. This is not an implementation detail – it is exactly how method binding works. When you access obj.method, Python does not find a pre-bound method stored somewhere. It calls function.__get__(obj, type(obj)) on the fly, which returns a bound method with the instance already attached.

Functions Define `get`


              123456
            
# Functions have __get__ – they are non-data descriptors
def greet(self):
    return f"Hello from {self.__class__.__name__}"

print(hasattr(greet, "__get__"))   # True
print(hasattr(greet, "__set__"))   # False – non-data descriptor

How Method Binding Works

When you write obj.method(), Python's attribute lookup finds method on the class (not the instance), sees it is a descriptor, and calls method.__get__(obj, type(obj)). This returns a bound method – a callable that prepends obj as the first argument:


              1234567891011121314151617181920
            
# Observing method binding step by step
class Analyst:
    def __init__(self, name):
        self.name = name

    def generate_report(self, period):
        return f"{self.name}: report for {period}"

analyst = Analyst("Alice")

# These are equivalent
print(analyst.generate_report("Q1"))

raw_function = Analyst.__dict__["generate_report"]
bound_method = raw_function.__get__(analyst, Analyst)
print(bound_method("Q1"))  # Same result

# The bound method carries the instance
print(bound_method.__self__ is analyst)  # True
print(bound_method.__func__ is raw_function)  # True

Why Functions Are Non-Data Descriptors

Because functions only define __get__ (not __set__), they are non-data descriptors. This means they can be shadowed by instance attributes with the same name:


              123456789101112
            
# Instance attribute shadows a method of the same name
class Processor:
    def run(self):
        return "Running processor"

processor = Processor()
print(processor.run())   # "Running processor" – calls method

# Shadowing the method with an instance attribute
processor.run = lambda: "Overridden at instance level"
print(processor.run())   # "Overridden at instance level" – instance wins
print(Processor.run(processor))  # Still accessible via the class

This is by design – it allows per-instance method overriding without subclassing.

Lambda and Partial as Descriptors

Any callable defined with def or lambda that lives in a class body is a non-data descriptor. functools.partial objects are not descriptors – they do not define __get__, so they do not bind:


              12345678910111213
            
import functools

# Comparing function binding vs partial (no binding)
def regular_method(self, value):
    return f"{self.__class__.__name__}: {value}"

class Report:
    method = regular_method                              # Binds on access
    partial_method = functools.partial(print, "prefix") # Does not bind

report = Report()
print(report.method("Q1"))          # "Report: Q1" – self is bound
print(report.partial_method("Q1"))  # "prefix Q1" – no self bound

The Full Picture

War alles klar?

Danke für Ihr Feedback!

Abschnitt 2. Kapitel 4

Fragen Sie AI

Fragen Sie alles oder probieren Sie eine der vorgeschlagenen Fragen, um unser Gespräch zu beginnen

Functions as Non-Data Descriptors

Functions Define `get`


              123456
            
# Functions have __get__ – they are non-data descriptors
def greet(self):
    return f"Hello from {self.__class__.__name__}"

print(hasattr(greet, "__get__"))   # True
print(hasattr(greet, "__set__"))   # False – non-data descriptor

How Method Binding Works


              1234567891011121314151617181920
            
# Observing method binding step by step
class Analyst:
    def __init__(self, name):
        self.name = name

    def generate_report(self, period):
        return f"{self.name}: report for {period}"

analyst = Analyst("Alice")

# These are equivalent
print(analyst.generate_report("Q1"))

raw_function = Analyst.__dict__["generate_report"]
bound_method = raw_function.__get__(analyst, Analyst)
print(bound_method("Q1"))  # Same result

# The bound method carries the instance
print(bound_method.__self__ is analyst)  # True
print(bound_method.__func__ is raw_function)  # True

Why Functions Are Non-Data Descriptors

Because functions only define __get__ (not __set__), they are non-data descriptors. This means they can be shadowed by instance attributes with the same name:


              123456789101112
            
# Instance attribute shadows a method of the same name
class Processor:
    def run(self):
        return "Running processor"

processor = Processor()
print(processor.run())   # "Running processor" – calls method

# Shadowing the method with an instance attribute
processor.run = lambda: "Overridden at instance level"
print(processor.run())   # "Overridden at instance level" – instance wins
print(Processor.run(processor))  # Still accessible via the class

This is by design – it allows per-instance method overriding without subclassing.

Lambda and Partial as Descriptors


              12345678910111213
            
import functools

# Comparing function binding vs partial (no binding)
def regular_method(self, value):
    return f"{self.__class__.__name__}: {value}"

class Report:
    method = regular_method                              # Binds on access
    partial_method = functools.partial(print, "prefix") # Does not bind

report = Report()
print(report.method("Q1"))          # "Report: Q1" – self is bound
print(report.partial_method("Q1"))  # "prefix Q1" – no self bound

The Full Picture

War alles klar?

Danke für Ihr Feedback!

Abschnitt 2. Kapitel 4

Functions as Non-Data Descriptors

Functions Define __get__

How Method Binding Works

Why Functions Are Non-Data Descriptors

Lambda and Partial as Descriptors

The Full Picture

Functions as Non-Data Descriptors

Functions Define __get__

How Method Binding Works

Why Functions Are Non-Data Descriptors

Lambda and Partial as Descriptors

The Full Picture

Functions Define `get`

Functions Define `get`