如何基于两个值比较数据（.csv）文件中的行，然后使用Python汇总数据？

Question

I have a .csv file that has 25 columns. In this data, column 18 is a People_ID, and column 19 is a Donation Date. I have pre-sorted the data using Linux so that all People ID's appear together, sorted by donation date in descending order.

Here is where I'm not sure how to proceed. I need to find all lines that have the same People_ID and Donation Date, sum various values, and then output a single line into the output. So essentially, every line in the file would be either a different customer, or a different donation date by the same customer. Would it be best to use a dictionary using the People_ID as the key? How would this look syntactically?

I was thinking something like this:

with open("file.csv") as csv_file:
for row in csv.reader(csv_file, delimiter=','):
    if row[18] in data_dict:
        # something something

Answer 1

Since you have pre-sorted, it could be organized to call a function once per person, each time called with the rows for a particular person.

Since the data is pre-sorted, we assume the rows for person 1 are together, then the rows for person 2 (or 39, or some other number), etc. So we need to detect when the person id in field 18 changes. To do this, we use a variable last_person to track which person we are processing. The variable row_cache will collect the rows for a single person.

def process_person(rows):
    if len(rows)==0:
         return
    # do something with the rows for this person
    # and print the result somewhere useful

last_person = 0
row_cache = []
with open("file.csv") as csv_file:
for row in csv.reader(csv_file, delimiter=','):
    if row[18]==last_person: 
        row_cache.append(row)
    else:
        process_person(row_cache)
        row_cache = [row]
        last_person = row[18]
process_person(row_cache)

Answer 2

I'd recommend an object-oriented approach.

import csv

class Transaction:
    def __init__(self, fields):
        self.name, self.age, self.car, self.ident = fields # whatever fields you have
        # keep in mind these are all strings,
        # so you may need to process them before analysis
    def calculation(self):
        return self.age + self.id

transactions = {}
with open('csv_file.csv', newline='') as f:
    for row in csv.reader(f):
        bucket = tuple(row[18:20])
        if bucket in transactions:
            transactions[bucket].append(Transaction(row))
        else:
            transactions[bucket] = [Transaction(row)]

for bucket in transactions:
    print(bucket, sum(item.amount for item in bucket.values()))

This defines a Transaction class, instances of which contain the various fields which will come from the CSV file. It then starts a dictionary of transactions and looks through the CSV file, adding a new Transaction object into a new bucket (if the given ID and date haven't been seen before) or into an existing bucket (if the given ID and date have been seen before).

Then it goes through this dictionary and performs a calculation for each bucket, printing the bucket and the result of the calculation.