CODEWITHSUNDEEP
python
Danish
Danish

Reputation: 3798

Initialize a Python Constant Using a Function Declared After It

I'm used to writing the following piece of code in Java or C# without errors. It allows me to centralize the conversion of string to a floating point representation. Unlike the simple implementation below, I have a lot more going on in there to handle commas or brackets for negative numbers in Excel.

I can understand why Python gives an error. Python code is not compiled but interpreted in sequence and toFloat does not exist for Python while its processing the declaration of ZERO .

I was wondering if there is a pythonic way of doing this without having to pull the toFloat function to another module OR without having to not declare ZERO at the very top?

ZERO = toFloat('0.0')    # <-- ERROR on this line "Undefined variable: toFloat"

def toFloat(val):
    """Function to convert a string to a float"""
    return Decimal(val)

Upvotes: 0

Views: 1453

Answers (3)

rlms
rlms

Reputation: 11060

Here is the Pythonic way of doing this:

def toFloat(val):
    """Function to convert a string to a float"""
    return Decimal(val)

ZERO = toFloat('0.0')

Preferring to define your constants before things they use just isn't how things work in Python. The best thing to do IMHO would be learning to like this way of doing things, rather than relying on clumsy workarounds. Python isn't like Java, not everything has to be wrapped in a load of classes. This often simplifies code a lot. As a general rule of thumb, if you make a class in Python, such as Tools in your example, that doesn't have a clear concrete purpose, you should probably take it out of the program.

Upvotes: 0

Danish
Danish

Reputation: 3798

I came up with the following piece of code, which I think is clean and keeps the declarations at the top while allowing me to keep the code below but still use it to initialize my variables.

class Tools(object):

    def __init__(self):
        self.ZERO = self.toFloat('0.0')

    def toFloat(self, val):
        return float(val)

tools=Tools()
print tools.ZERO

Upvotes: 0

Hyperboreus
Hyperboreus

Reputation: 32449

You are accessing an undefined symbol, which is a bad idea in many languages. Python has no idea, what the symbol toFloat might refer to.

For illustration, what would you expect this code to do? Should ZERO be zero or one?

ZERO = toFloat('0.0')

def toFloat(val): return Decimal(val)
def toFloat(val): return 1

Or in this case:

ZERO = toFloat('0.0')

if ZERO: from libone import toFloat #where toFloat returns 0
else: from libtwo import toFloat #where toFloat returns 1

The thing is you can redefine your symbols as many times you want. E.g. this is legit code:

def f(): return 0
print(f())
def f(): return 1
print(f())

If it is of vital importance that ZERO be declared at the top of your code, maybe you can use this ugly hack for your constants:

Put this in another file, let's say const.py:

class Constant:
    def __init__(self, cf):
        self.cf = cf

    @property
    def c(self):
        try: return self.__c
        except: self.__c = self.cf()
        return self.__c

And then in your file, you can use:

from const import Constant

ZERO = Constant(lambda: toFloat('0.0'))
def toFloat(x): return 42
print(ZERO.c)

Upvotes: 1

Related Questions