Python: Copying named tuples with same attributes / fields

Question

I am writing a function that takes a named tuple and must return a super set of that tuple.

For example if I was to receive a named tuple like this:

Person(name='Bob', age=30, gender='male')

I want to return a tuple that looks like this:

Person(name='Bob', age=30, gender='male', x=0)

Currently I am doing this:

tuple_fields = other_tuple[0]._fields
tuple_fields = tuple_fields + ('x')
new_tuple = namedtuple('new_tuple', tuple_fields)

Which is fine, but I do not want to have to copy each field like this:

tuple = new_tuple(name=other_tuple.name, 
                  age=other_tuple.age, 
                  gender=other_tuple.gender, 
                  x=0)

I would like to be able to just iterate through each object in the FIRST object and copy those over. My actual tuple is 30 fields.

Departed

Possible duplicate of Inherit namedtuple from a base class in python – Cinderellacindi

Answer 1

You could try utilizing dict unpacking to make it shorter, eg:

tuple = new_tuple(x=0, **other_tuple._asdict())

Carlin

Whether namedtuples have a __dict__ and how it works is rather inconsistent across Python versions. I don't think the latest 3.5 revision has it. Having a __dict__ for namedtuples was a bad idea in the first place, and it caused some nasty bugs. – Onepiece

If you want a dict representation of a namedtuple, there's the _asdict method. – Onepiece

@Onepiece Thanks, I'll update my answer for that. __dict__ worked fine for me, but I'm stuck on Python 3.4. – Carlin

Answer 2

The accepted version is broken, you will get a new_tuple() got multiple values for argument 'x' error.

You should use _replace instead:

https://docs.python.org/3/library/collections.html#collections.somenamedtuple._replace

Noctambulous