哪段儿代码更优雅，可读性更高，更Pythonic？

哪段儿代码更优雅，可读性更高，更Pythonic？

题目要求

找出一个锯齿数组里长度大于5的子数组

在符合要求的子数组里的数据里找出所有偶数

如果数据小于10的话乘以2,大于10的除以2

最后统计符合要求的数据的和

测试数据

inputdata = [
       [2,8,9,13,72,67,88,35,44],
       [33,28,47,2,10,45,66,92],
       [22,34,60,43,0,72,52],
       [10,11,53,58]
        ]
print sum1(inputdata)

@蛙蛙王子 python

https://github.com/onlytiancai/codesnip/blob/master/python/goodorbad.py

def sum1(input):
    return sum((lambda x: x < 10 and x*2 or x/2)(num)
            for seq in inputdata
            if len(seq) >= 5
            for num in seq
            if num % 2 == 0)
def sum2(input):
    def getsublist():
        for sublist in input:
            if len(sublist) >= 5:
                yield sublist
    def filterdata(sublist):
        for data in sublist:
            if data % 2 == 0:
                yield data
    def processdata(data):
        if data < 10:
            return data * 2
        return data / 2
    result = 0
    for sublist in getsublist():
        for data in filterdata(sublist):
            result += processdata(data)
    return result
def sum3(input):
    return sum(num*2 if num < 10 else num/2
            for seq in inputdata
            if len(seq) >= 5
            for num in seq
            if num % 2 == 0)

@pw_网易有道 python

https://gist.github.com/1891339

def sum4(input):
    '''
    '''
    def filterList(matrix):
        return filter(lambda x: len(x) >= 5, matrix)
    def filterData(list):
        return filter(lambda x: x % 2 is 0, list)
    def processData(data):
        return data < 10 and data * 2 or data / 2
    def reduceList(list):
        return reduce(lambda x, y: x + processData(y), filterData(list), 0)
    def reduceMatrix(matrix):
        return reduce(lambda result, list: result + reduceList(list), filterList(matrix), 0)
    return reduceMatrix(input)

@w1e3 python

sum(map(lambda x: x/((cmp(x, 10)+1) or 0.5),
    itertools.chain(*[filter(lambda x: not x%2, s)
    for s in inputdata if len(s) > 5])))

@Leksah Haskell

sum [ if y < 10 then y*y else y | x <-inputdata, length(x) > 5 , y <- x , y `mod` 2 ==0 ] i

@codeplayer Haskell

sum' = sum
       . map (\x -> if x<10 then x*x else x)
       . filter ((==0) . (`mod` 2))
       . concat
       . filter ((>5) . length)

@时蝇喜箭 Ruby

inputdata.select{|seq| seq.length > 5}
    .map{|seq| seq.select{|num| num % 2 == 0}.map{|num| num < 10? num*2 : num/2}}
    .flatten.inject(&:+)

@朴灵 Javascript&Underscore

_.chain(inputdata)
    .filter(function (arr) { return arr.length > 5; })
    .map(function (arr) { return _.map(arr, function (num) { return num < 10 ? num * 2 : num; }); })
    .flatten()
    .reduce(function (memo, num) { return memo + num; }, 0)
    .value();