typename in collections.namedtuple

[Skaperen]

does anyone know what collections.namedtuple's first argument, docs refer to as typename, is actually used for? i'm trying to figure out what i should put there.

[buran]

So to say - the name of the tuple, i.e. the typename

from collections import namedtuple

Foo = namedtuple('Bar', 'foo bar')
eggs = Foo(1, 3)
print(eggs)

Output:
Bar(foo=1, bar=3)

as you can see Foo is used for the instantiation, but the name (Bar) is the one provided as typename

[Skaperen]

gotcha, so it is creating a type. at a minimum each named tuple type name needs to be unique? does this namespace also need to avoid typenames of non-namedtuple types like "float"?

[buran]

Don't know the answers for sure but lets check:

from collections import namedtuple
Spam = namedtuple('float', 'foo bar')
eggs = Spam(1, 2)
print(eggs)
print(type(eggs))
x = 1.2
print(type(x))
print(float('1.5'))

Spam2 = namedtuple('float', 'a, b')
eggs2 = Spam2(3, 4)
print(eggs)
print(eggs2)

Output:
float(foo=1, bar=2)
<class '__main__.float'>
<class 'float'>
1.5
(sbox) boyan@buranhome ~/sandbox2 $ /home/boyan/sandbox2/sbox/bin/python /home/boyan/sandbox2/enemy.py
float(foo=1, bar=2)
<class '__main__.float'>
<class 'float'>
1.5
float(foo=1, bar=2)
float(a=3, b=4)

It works but I would say it's confusing (both using built-in type as typename or using non-unique typename)

[Skaperen]

what i would wonder is if there is any interference or collision with the the real float type.

---

and i have no plan to do other than use unique meaningful type names. i am looking over old functions to see what might get benefit from a named tuple.