一般使用lambda函数作为字符串过滤Python对象数组中的值

Question

I'm working on a generic framework, and at some point I'm trying to filter variables in a generic manner. Take for example the following class:

class X:
    def __init__(self, val):
        self.val = val

Then I have an array of objects with arbitrary val values:

arr = [X("1"), X("2"), X("2"), X("3"), X("foo"), X("5"), X("3")]

My goal is to have a function which takes a lambda function and a variable names as strings (because they are user provided), which is currently implemented as follows:

def process(lambda_f_str, var_str):
    # should return [1, 2, 3, 5]
    return list(set(filter(lambda x: x, map(eval(lambda_f_str), eval(var_str)))))

I want to be able to call this method with a given lambda function string that will return me only the unique integer values of these objects (In this example I'd expect [1 2 3 5] , the order doesn't matter).

I've tried something like this:

process("lambda x: x.val if isinstance(x.val, int) else None", "arr")

But this doesn't work since the integers are still passed as strings in my array (and I have no control over that, this is user-provided so I can't make any assumption).

I wanted to do the equivalent of

try:
    int(x)
except:
    # do something if not an int

But I don't think you can do that in a lambda function ... can you? At least I haven't found out how.

I'd be interested to see how I can write this lambda to do this filtering in a generic manner. I'd rather avoid changing the process method since this is a generic function which does other things not included in this sample code for the sake of brevity.

Answer 1

Try this...

s = "lambda x: int(x.val) if x.val.isdigit( ) else None"
print process(s, "arr")

I get back...

[1, 2, 3, 5]

Note that the .isdigit( ) method will fail if the numbers are given with commas or decimals.

Answer 2

Do you really need the evals?

f = lambda x: x.val if isinstance(x.val, int) else (int(x.val) if isinstance(x.val, basestring) and x.isdigit() else None)

Not 100% proof, eg it doesn't take 123L which is valid Python.