Question

I have 2 list containing dictionaries as follows:

listone = [{'unit1': {'test1': 10}}, 
           {'unit1': {'test2': 45'}, 
           {'unit2': {'test1': 78'}, 
           {'unit2': {'test2': 2'}}]

listtwo = [{'unit1': {'test1': 56}}, 
           {'unit1': {'test2': 34'}, 
           {'unit2': {'test1': 23'}, 
           {'unit2': {'test2': 5'}}]

I also do have all the unit names & test-names in separate lists:

units = ['unit1', 'unit2']
testnames = ['test1,'test2']

How could I find the delta for each test value, i.e. val of (test2 – test1), so that I could finally arrange the data as follows:

unit1, test1, delta
unit1, test2, delta
unit2, test1, delta
unit2, test2, delta

So far, I have these:

def delta(array1, array2):
        temp = []
        temp2 = []
        tmp = []
        tmp2 = []
        delta = []
        for unit in units:
            for mkey in array1:
                for skey in mkey:
                    if skey == unit:
                        temp.append(mkey[skey])
                        floater(temp) #floats all the values
                        for i in testnames:
                            for u in temp:
                                tmp.append(u[i])
                                tmp = filter(None, tmp2)

            for mkey in array2:
                for skey in mkey:
                    if skey == unit:
                        temp.append(mkey[skey])
                        floater(temp2)
                        for i in testnames:
                            for u in temp2:
                                tmp2.append(u[i])
                                tmp2 = filter(None, tmp2)

        delta = [tmp2 - tmp for tmp2, tmp in zip(tmp2, tmp)] 
        print delta

delta(listone,listtwo)

Unfortunately, the code gives Keyerror. 🙁
Help, please. Thanks.

Answer 1

Similar but a bit more encapsulated:

from collections import defaultdict

listone = [
    {'unit1': {'test1': 10}},
    {'unit1': {'test2': 45}}, 
    {'unit2': {'test1': 78}}, 
    {'unit2': {'test2': 2}}
]

listtwo = [
    {'unit1': {'test1': 56}},
    {'unit1': {'test2': 34}}, 
    {'unit2': {'test1': 23}}, 
    {'unit2': {'test2': 5}}
]

def dictify(lst):
    res = defaultdict(lambda: defaultdict(int))
    for entry in lst:
        for unit,testentry in entry.iteritems():
            for test,val in testentry.iteritems():
                res[unit][test] = val
    return res
    # returns dict['unitX']['testY'] = val

def genDeltas(dictA, dictB):
    units = dictA.keys()
    units.sort()
    tests = dictA[units[0]].keys()
    tests.sort()
    for unit in units:
        _A = dictA[unit]
        _B = dictB[unit]
        for test in tests:
            yield unit,test,(_B[test]-_A[test])

for unit,test,delta in genDeltas(dictify(listone),dictify(listtwo)):
    print "{0}, {1}, {2}".format(unit,test,delta)

Edit: to find field-averages:

class Avg(object):
    def __init__(self, total=0.0, num=0):
        super(Avg,self).__init__()
        self.total = total
        self.num   = num

    def add(self, value):
        self.total += value
        self.num   += 1

    def value(self):
        return self.total / self.num

def avgBy(data, field=0):
    res = defaultdict(Avg)
    for unit,testdict in data.iteritems():
        for test,val in testdict.iteritems():
            res[(unit,test)[field]].add(val)
    return {item:avg.value() for item,avg in res.iteritems()}

dictone = dictify(listone)
avg_by_unit = avgBy(dictone, 0)
print(avg_by_unit)
avg_by_test = avgBy(dictone, 1)
print(avg_by_test)