pingul pingul - 1 year ago 54
Python Question

How to overload * argument unpacking operator?

I have data like

data = [[t1, t2, ...], [v1, v2, ...]]
. I want to wrap this in a class so I can call
instead of having to use

I tried to do this with the following:

class Variable:
def __init__(self, data):
self.t = data[0]
self.v = data[1]

def __getitem__(self, key):
if key == 0:
return self.t
elif key == 1:
return self.v
raise ValueError("not valid key '{}'".format(key))

def __setitem__(self, key, value):
if key == 0:
self.t = value
elif key == 1:
self.v = value
raise ValueError("not valid key '{}'".format(key))

The reason for the
overloading is for backwards compability so that
still works. This works for most things, however I run into problems with the following call:

func_that_takes_two_arguments(*data) # unpacking data

The error I get is

/Users/pingul/Workspace/lhcfill/ in __getitem__(self, key)
52 return self.val
53 else:
---> 54 raise ValueError("not valid key '{}'".format(key))
56 def __setitem__(self, key, value):
ValueError: not valid key '2'

How can I make my class work properly with the argument unpacking operator?

Answer Source

The * operator works by iterating over the object. This iteration can well be performed with only implementing __getitem__(), but your implementation is faulty. Instead if raising ValueError, you should throw IndexError which signals the end of the iteration.

See also which explicitly states

Note: for loops expect that an IndexError will be raised for illegal indexes to allow proper detection of the end of the sequence. states that this is called the "sequence protocol".