CSV文件的Python计数器

Question

Im new to Python, and I need some help to get the results of a survey. I have a CSV file, which looks like this:

Person, Gender, Q1, Q2, Q3
professor, male, agree, not agree, agree
professor, male, agree, agree, agree
professor, female, neutral, not agree, agree
Professor, female, agree, agree, agree
student, female, agree, not agree, not agree
student, female, no answer, not agree, agree
student, male, no answer, no answer, agree

I want to count the number of times the different answers occur per person and gender. For example Q1: (professor, male: agree, 2), (professor, female: agree 1; neutral 1) and so on. I have tried this so far:

import csv
from collections import Counter
with open('survey.csv') as csvfile:
    reader = csv.reader(csvfile, delimiter=',', dialect = csv.excel_tab)
    counts = Counter(map(tuple,reader))
    print [row for row in reader if row]
    print list(csv.reader(csvfile))

But I think because I have only strings, I do not get any result. Moreover, I still don't know how to get the data by people/gender. Thanks a lot in advance!

Answer 1

Using pandas you could do something like:

import pandas as pd
my_data = pd.read_csv('survey.csv')
# To summarize the dataframe for everything together:
print my_data.describe()
print my_data.sum()

# To group by gender, etc.
my_data.groupby('Gender').count()

Answer 2

If you don't want to switch to pandas, you need to do a bit of analysis on the rows after you read them. Something like the following (untested). This uses Counter objects which behave a lot like ordinary dicts except that referring to a key that does not (yet) exist creates it automagically and gives it the value 0, rather than raising KeyError .

from collections import Counter

counters = []
for row in reader:
   for colno,datum in enumerate(row):
       if colno >= len(counters): # do we have a counter for this column yet?
           counters.append( Counter() ) # if not, add another Counter
       counters[colno][datum] += 1

for counter in counters:
    print(counter)

If the first row of your csv file is some column headers, you can read it in advance and then use it to annotate the list of counters. I'll leave formatting the contents of the counters prettily to you as an exercise, should raw dumps of counter objects be deemed too ugly.