Python-在字典元素中查找平均值

Question

I have dict like:

dict = [{'a':2, 'b':3}, {'b':4}, {'a':1, 'c':5}]

I need to get average of all different keys. Result should looks like:

avg = [{'a':1.5, 'b':3.5, 'c':5}]

I can get summary of all keys, but Im failing to realize how can I count same keys in order to get average number.

Answer 1

This can be easily done with pandas :

>>> import pandas
>>> df = pandas.DataFrame([{'a':2, 'b':3}, {'b':4}, {'a':1, 'c':5}])
>>> df.mean()
a    1.5
b    3.5
c    5.0
dtype: float64

If you need a dictionary as result:

>>> dict(df.mean())
{'a': 1.5, 'b': 3.5, 'c': 5.0}

Answer 2

You could create an intermediate dictionary that collects all encountered values as lists:

dct = [{'a':2, 'b':3}, {'b':4}, {'a':1, 'c':5}]
from collections import defaultdict
intermediate = defaultdict(list)

for subdict in dct:
    for key, value in subdict.items():
        intermediate[key].append(value)

# intermediate is now: defaultdict(list, {'a': [2, 1], 'b': [3, 4], 'c': [5]})

And finally calculate the average by dividing the sum of each list by the length of each list:

for key, value in intermediate.items():
    print(key, sum(value)/len(value))

which prints:

b 3.5
c 5.0
a 1.5

Answer 3

You can use a for loop with a counter and then divide the sum of each by the counter.

Also it is weird you are calling the array/list a dict...

I'd suggest something like this:

Create a new dict: letter_count = {}

-For loop over the current dicts

-Add the letter to the letter count if it doesn't exist

-If it does exist, update the value with the value of the item (+=number) as well as update the counter by one

-Once the for loop is done, divide each value by the counter

-Return the new dict letter_count

Answer 4

I thought of adding a unique answer using PyFunctional

from functional import seq

l = [{'a':2, 'b':3}, {'b':4}, {'a':1, 'c':5}]
a = (seq(l)
     # convert dictionary to list
     .map(lambda d: seq(d).map(lambda k: (k, d[k])))
     .flatten()
     # append 1 for counter
     .map(lambda (k, v): (k, (v, 1)))
     # sum of values, and counts
     .reduce_by_key(lambda a, b: (a[0]+b[0], a[1]+b[1]))
     # average
     .map(lambda (k, (v, c)): (k, float(v)/c))
     # convert to dict
     .to_dict()
)
print(a)

Output

{'a': 1.5, 'c': 5.0, 'b': 3.5}