If a runtime error occurs in a function, Python checks whether an exception handler has been created in the function to handle this error. If no exception handler is defined inside the function, Python checks whether an exception handler is defined in the code that called the function.

Here is an example that illustrates this concept:

def divide_x_y(x, y):
    try:
        z = x / y
        return z
    except TypeError, errmsg:
        print "Type error:", errmsg

try:   
    result = divide_x_y(23, 0)
except ZeroDivisionError:
    print "Division by zero error in function"

In this example, the function divide_x_y() detects type mismatch errors, but ignores division by zero errors. The code that calls divide_x_y() contains its own exception handler; this handler detects the division by zero and displays an error message.

If an error is detected both inside and outside a function, the exception handler inside the function is used:

def divide_x_y(x, y):
    try:
        z = x / y
        return z
    except TypeError, errmsg:
        print "Type error:", errmsg
    except ZeroDivisionError:
        print "Division by zero error"

try:   
    result = divide_x_y(23, 0)
except ZeroDivisionError:
    print "This message will never be printed"

Here, the division by zero error is handled inside divide_x_y(). This means that the exception handler outside the function will never be executed. If you want to handle an error both inside and outside a function, use the raise statement:

def divide_x_y(x, y):
    try:
        z = x / y
        return z
    except TypeError, errmsg:
        print "Type error:", errmsg
    except ZeroDivisionError:
        print "Division by zero error"
        raise

try:   
    result = divide_x_y(23, 0)
except ZeroDivisionError:
    print "This message is now printed"

The raise statement regenerates the division by zero error and passes it up to the code that called the divide_x_y() function. This means that both error messages are displayed.

	You can use raise to generate any exception you want. For example, the following code generates a TypeError exception: def generate_type_error(): raise TypeError, "This is a generated error." try: generate_type_error() except TypeError, errmsg: print "TypeError detected:", errmsg The parameters to raise specify the error you want to generate and the information you want to provide with the error. This code produces the following output: TypeError detected: This is a generated error. You can also create your own custom errors and use raise to generate them. For more information, see the User-defined Exceptions section of the Python Tutorial guide.