Python 字典的具体值平均值

Question

I have a dictionary:

{201001: '-28.3', 201002: '-23.8', 201003: '-24.2', 201004: '-14.1',
 201005: '-7.5', 201006: '0.1', 201007: '5.0', 201008: '5.6',
 201009: '2.6', 201010: '-5.0', 201011: '-12.1', 201012: '-23.6',
 201101: '-23.0', 201102: '-21.7', 201103: '-21.9', 201104: '-19.5',
 201105: '-6.0', 201106: '0.7', 201107: '4.8', 201108: '5.2',
 201109: '2.3', 201110: '-4.9', 201111: '-18.5', 201112: '-23.8'}

This dictionary represents the temperature of Alaska. 201001: '-28.3' represent temperature of Alaska in 2010 January is -28.3 celsius.)

I want to get the average temperature of winter year by year which is

( January's + February's ) / 2 .

For example 2010's winter average temperature is (-28.3 + -23.8) / 2 .

I think seeing and directly typing average year by year isn't elegant because there are more years 2000~2019.

(I typed only 2 years 2010,2011 for readability.) How to get average of specific data year by year in dictionary?

Answer 1

Just a simple loop should do it, no?

from typing import Dict
def avg_winter_temps(temps: Dict[int, float]) -> Dict[int, float]:
    winter_temps = {}
    for year in range(2010, 2019+1):
        jan, feb = year*100 + 1, year*100 + 2
        winter_temps[year] = (temps[jan] + temps[feb])/2
    return winter_temps

Answer 2

Temperature by Year

by_year = {}
for k, v in d.items():
  # k // 100 is year
  by_year.setdefault(k //100, []).append(float(v))

Average Temperature in Winter

avg_winter = {}
for k, months in by_year.items():
    # months[:2] to average first two months
    avg_winter[k] = sum(months[:2])/len(months[:2])

Output

from pprint import pprint
pprint(by_year)
pprint(avg_winter)

by_year

{2010: [-28.3,
        -23.8,
        -24.2,
        -14.1,
        -7.5,
        0.1,
        5.0,
        5.6,
        2.6,
        -5.0,
        -12.1,
        -23.6],
 2011: [-23.0,
        -21.7,
        -21.9,
        -19.5,
        -6.0,
        0.7,
        4.8,
        5.2,
        2.3,
        -4.9,
        -18.5,
        -23.8]}

avg_winter

{2010: -26.05, 2011: -22.35}

Answer 3

You can use regular expressions to find keys that end in "01" or "02":

import re
dictionary = {201001: '-28.3', 201002: '-23.8', 201003: '-24.2', 201004: '-14.1',
 201005: '-7.5', 201006: '0.1', 201007: '5.0', 201008: '5.6',
 201009: '2.6', 201010: '-5.0', 201011: '-12.1', 201012: '-23.6',
 201101: '-23.0', 201102: '-21.7', 201103: '-21.9', 201104: '-19.5',
 201105: '-6.0', 201106: '0.7', 201107: '4.8', 201108: '5.2',
 201109: '2.3', 201110: '-4.9', 201111: '-18.5', 201112: '-23.8'}
key_list = dictionary.keys()
months = 0
sum = 0
averages = []
for key in key_list:
    match = re.search("01$|02$", str(key)) # String ends in '01; or in '02'
    if match:
        if months == 0:
            sum += float(dictionary.get(key))
            months += 1
        else:
            sum += float(dictionary.get(key))
            months = 0
            averages.append(sum / 2)
            sum = 0
print(averages)

Output:

[-26.05, -22.35]

Answer 4

This is not intended as an answer to the question asked, but to supplement the existing answers if you want to take into account that to calculate a time average of temperature over the January-February period the monthly averages should be weighted by the lengths of the months concerned. The function calc_jan_feb_average can be used to calculate this weighted average. It requires the year as an input parameter because of leap years.

import datetime

def get_length_feb(year):
    return (datetime.datetime(year, 3, 1)
            - datetime.datetime(year, 2, 1)).total_seconds()

def calc_jan_feb_average(year, jan_temp, feb_temp):
    length_jan = 31 * 86400
    length_feb = get_length_feb(year)

    return (
        ((jan_temp * length_jan) + (feb_temp * length_feb)) /
        (length_jan + length_feb)
    )

Answer 5

Is this what you're looking for?

temps = {201001: '-28.3', 201002: '-23.8', 201003: '-24.2', 201004: '-14.1',
         201005: '-7.5', 201006: '0.1', 201007: '5.0', 201008: '5.6',
         201009: '2.6', 201010: '-5.0', 201011: '-12.1', 201012: '-23.6',
         201101: '-23.0', 201102: '-21.7', 201103: '-21.9', 201104: '-19.5',
         201105: '-6.0', 201106: '0.7', 201107: '4.8', 201108: '5.2',
         201109: '2.3', 201110: '-4.9', 201111: '-18.5', 201112: '-23.8'}

avgs = {}
jan = min(temps.keys()) // 100 * 100 + 1
limit = max(temps.keys())
while jan <= limit:
    jt = temps.get(jan)
    feb = jan + 1
    ft = temps.get(feb)
    if jt is not None and ft is not None:
        avgs[f'{jan} - {feb}'] = (float(jt) + float(ft)) / 2
    jan += 100

for period, avg in avgs.items():
    print(f'{period} = {avg}')

Output:

201001 - 201002 = -26.05
201101 - 201102 = -22.35