First, you need to get the csv fields by splitting on the actual delimiter you're using (which is a semicolon, not a comma):
csv.reader(csvfile, delimiter=';')
The result of iterating on reader now will be a series of lists of strings, looking like the following:
['21/10/2017 0:00', '123,85', '88,8']
(Note: this is the python representation for strings, the ' characters are not actually part of the data)
Now, to get to the actual numbers, you need to turn those strings to values.
The second and the third are more or less straightforward, but you need to take care of that comma. The value you have is using a locale in which decimal values are separated by a comma, python expects a dot. So, we can convert them as follows (Let line be one line from the reader):
second_number = float(l[1].replace(',','.'))
third_number = float(l[2].replace(',','.'))
For the date the thing is more complicated. Assuming you're just interested in the numbers in the date and not a full conversion to some datetime value, this is what you could do:
date, time = line[0].split(' ') #separate "21/10/2017" from "0:00"
day, month, year = [int(v) for v in date.split('/')]
hour, minute = [int(v) for v in time.split(':')]
I hope this is clear enough and matching what you need
csvmodule will do the conversion for you. Which, I assume, is the reason you used that package in the first place['21/10/2017 0:00', '123,85', '88,8']?['21/10/2017 0:00;123,85;88,8']They maybe meant:['21/10/2017', '0:00;123,85;88,8']