Question

I have two lists of fixed-length tuples. This function calculates a fraction (ratio) for respective elements (note, fX() does not use comprehension for readability here).

>>> def fX(a,b):  
>>>    c=[]  
>>>    for i in range(len(a)):  
>>>        c.append([a[i][x]/float(a[i][x]+b[i][x]) for x in range(len(a[i]))])
>>>    return c  

When all values are non-zero, fX() works:

>>> a[0]=(3, 4, 17, 9.6667, 6.6583, 0.4310, 1)  
>>> b[0]=(4, 4, 12, 8.0, 3.2660, 0.0002, 1)  
>>> fX(a,b)  
>>> [[0.4286, 0.5, 0.5862, 0.5472, 0.6710, 0.9995, 0.5]]  

However, when any pair’s values sum to zeros, fX() fails:

>>> a[0]=(3, 4, 17, 9.6667, 6.6583, 0.4310, 0)  
>>> b[0]=(4, 4, 12, 8.0, 3.2660, 0.0002, 0)  
>>> fX(a,b)  
Traceback (most recent call last):  
  File "<pyshell#59>", line 1, in <module>  
    fX(a,b)  
  File "<pyshell#52>", line 4, in fX  
    c.append([a[i][x]/float(a[i][x]+b[i][x]) for x in range(len(a[i]))])  
ZeroDivisionError: float division  

I’m need a function, fY(), that gives this desired outcome without resorting to a brute force test of each value:

>>> a[i]=(3, 4, 17, 9.6667, 6.6583, 0.4310, 0)  
>>> b[i]=(4, 4, 12, 8.0, 3.2660, 0.0002, 0)  
>>> fY()  
>>> [[0.4286, 0.5, 0.5862, 0.5472, 0.6710, 0.9995, 0.0]]  

Thanks.

Answer 1

Use the a if x else b ternary operator (equivalent to the C/C++/Java expression x ? a : b) to put the conditional inside the list comprehension. This gives an efficient, Pythonic implementation:

def fY(a, b):
    return [[aij/float(aij + bij) if aij+bij != 0 else 0 for aij, bij in zip(ai, bi)]
            for ai, bi in zip(a, b)]