Dict key selected / set value selected are different when code is executed in shell and via script-CodePudding

Python 3.7.4 (default, Sep 8 2020, 19:45:30) [GCC 7.5.0] on linux

first_tuple = (1, 2, )
second_tuple = (1, 2, )
first_list = [1, 2, 3]
second_list = [1, 2, 3]


def main():
    # Question 1
    my_dict = {
        first_tuple: first_list,
        second_tuple: second_list,
    }
    print('dictionary length = {}'.format(len(my_dict)))
    for key in my_dict.keys():
        if id(key) == id(second_tuple):
            print("key selected is 'second_tuple'")
        else:
            print("key selected is 'first_tuple'")
        if id(my_dict[key]) == id(second_list):
            print("value selected is 'second_list'")
        else:
            print("key selected is 'first_list'")
    # Question 2`
    my_set = {first_tuple, second_tuple}
    print('set length = {}'.format(len(my_set)))
    if id(my_set.pop()) == id(second_tuple):
        print("'second_tuple' is considered")
    else:
        print("'first_tuple' is considered")

main()

Above snippet when executed in a python shell is giving output as:

dictionary length = 1
key selected is 'first_tuple'
value selected is 'second_list'
set length = 1
'first_tuple' is considered

And when same is executed as a script eg. python3.7 example.py

dictionary length = 1
key selected is 'second_tuple'
value selected is 'second_list'
set length = 1
'second_tuple' is considered

Why there is a difference? Is compiler doing some optimisation?

CodePudding user response：

Yes, it's a compiler optimization. The same tuple is being used for first_tuple and second_tuple when the script is compiled. This is allowed because tuples are immutable, so there's no need to distinguish between equal tuples.

You can see this with a much simpler example:

first_tuple = (1, 2, )
second_tuple = (1, 2, )
print(first_tuple is second_tuple)

This prints True in a script, but False in the shell.