如何使用Python中的多列计算CSV文件中日期之间的平均时间？

Question

I have a csv file with three different columns, namely Year, Month and Day. If I printed it out, it will be something like below:

csv_reader = [['2016', '6', '22'], ['2016', '10', '2'], ['2016', '11', '1'], ['2016', '11', '3'], ['2016', '11', '3'], ['2016', '11', '17'], ['2016', '11', '17'], ['2016', '11', '17'], ['2016', '12', '2'], ['2016', '12', '12'], ['2016', '12', '22'], ['2016', '12', '22'], ['2017', '1', '11'], ['2017', '3', '11'], ['2017', '3', '11'], ['2017', '5', '12'], ['2017', '5', '12'], ['2017', '5', '12']]

So both the csv file and the row are lists. I want to calculate the average days between each date from this dataset. I try to use:

from datetime import date
for value in csv_reader:
    dates = date(int(value))
    differences = [(dates[i]-dates[i-1]).days for i in range(1, len(dates))]
print(float(sum(differences))/len(differences))

But it returned an error stating that

TypeError: int() argument must be a string, a bytes-like object or a number, not 'list'

and I suspect if my code is actually correct. The expected result should calculate the sum of difference between each dates, and then divided by the total differences, which would return the average days. So, could any of you give advice on how to perform this task?

Answer 1

So we have

csv_reader = [['2016', '6', '22'], ['2016', '10', '2'], ['2016', '11', '1'], ['2016', '11', '3'], ['2016', '11', '3'], ['2016', '11', '17'], ['2016', '11', '17'], ['2016', '11', '17'], ['2016', '12', '2'], ['2016', '12', '12'], ['2016', '12', '22'], ['2016', '12', '22'], ['2017', '1', '11'], ['2017', '3', '11'], ['2017', '3', '11'], ['2017', '5', '12'], ['2017', '5', '12'], ['2017', '5', '12']]

first, to get a valid date from lets say, the first item, you need to covert the str to int, and only then call date :

date(*[int(d) for d in csv_reader[0]])

And you need to have a date instance for each 2 dates and subtract them:

date(*[int(d) for d in csv_reader[0]]) - date(*[int(d) for d in csv_reader[1]])

Afterwards, you get a timedelta object, which has .days . Sometimes it'll be negative, so you'll need to use the absolute value, with abs .

After you sum all those days, you need to do the average, relative to the number of elements.

All in all, this is the loop you need:

total = 0
for i in range(len(csv_reader)-1):  
    total += abs((date(*[int(d) for d in csv_reader[i]]) - date(*[int(d) for d in csv_reader[i+1]])).days)

>>> total
324
>>> total / len(csv_reader)
18

Answer 2

1. Convert list of str to the list of dates

    dates = [datetime.date(*[int(d) for d in ds]) for ds in sortedtime] 

2. Zip two lists of dates with shift for 1 day and calculate timedelta of days:

    delta_days = [(d_t[0] - d_t[1]).days for d_t in list(zip(dates[1:],dates))] 

3. The Average will be a simple operation

    avg_days = sum(delta_days)/len(delta_days) 

Answer 3

You can't just convert this string '2016,6,22' to int. You should delete commas before you pass a string into the int() function

Answer 4

尝试将date(int(value))更改为date(int(value.replace(',',''))) 。