Python CSV比较与乘法

Question

I want to double-check my logic on how to put this together in Python, so examples are appreciated.

I need to compare 2 CSV files (same format exactly with 2 rows and 6 columns) and provide the difference.

I need to pull both in and multiply row 2, columns 2-6 by specific values (5), total them separately, and then compare with each other (CSV2 total/CSV1 total), and present in a percentage format.

import csv and reader seem like the way to go, but the tricky part for me has been pulling it into a list I can multiply against different values (or should I use a collection?), and then comparing the two in the most concise/efficient manner.

Code update (based on 2nd answer- was great, thanks! but now encountering error with calling my row values integers):

import csv
file1 = open('csv1.csv', 'rb')
csv1 = csv.DictReader(file1)

file2 = open('csv2.csv', 'rb')
csv2 = csv.DictReader(file2)


myList = csv2.fieldnames
myList.append('Difference')

outFile = open('outFilename.csv', 'wb')
outCsv = csv.DictWriter(outFile, myList)

file1Dict = dict()
file2Dict = dict()

for row in file1:
    file1Dict[row['key value']]['Total1'] = {'Total1':(int(row[1]) * .75 + int(row[2]) * 2.25 + int(row[3]) * 3.5 + int(row[4]) * 5 + int(row[5]) * 25)}

for row in file2:
    file2Dict[row['key value']]['Total2'] = {'Total2':(int(row[1]) * .75, int(row[2]) * 2.25, int(row[3]) * 3.5, int(row[4]) * 5, int(row[5]) * 25)}

outFile.writeheader()

for stuff in file1Dict:
    file1Dict[stuff]['Difference'] = str(int(int(file1Dict[stuff]['Total2']) / int(file1Dict[stuff]['Total1'])) * 100) + '\%'
    outFile.writerow(file1Dict[stuff])

print 'difference'

Answer 1

I think you should use Python Pandas and the built-in read_csv function, since this will be very efficient and put it into a rectangular form where any kind of math operations are trivial to apply and easy to compare across two different imported data sets.

Note that after importing pandas , there is a global level read_csv , called just as pandas.read_csv("/path/to/file.csv") rather than needing to go through io.parsers as in the linked doc page above.

There's nothing wrong with doing this through standard modules; I just think if you plan to do aggregate or broadcasted math operations, the rectangular array provided by Pandas that uses the efficient NumPy math operations is the best choice.

Answer 2

import csv
file1 = open('filename1.csv', 'rb')
csv1 = csv.DictReader(file1)

file2 = open('filename2.csv', 'rb')
csv2 = csv.DictReader(file2)


myList = csv2.fieldnames
myList.append('Total1','Total2', 'Difference')

outFile = open('outFilename.csv', 'wb')
outCsv = csv.DictWriter(outFile, myList)

file1Dict = dict()
file2Dict = dict()

for rows in file1:
    file1Dict[rows['key value']] = {rows[0], rows[1], 'Total1':int(rows[1]) * 5}

for rows in file2:
    file1Dict[rows['key value']]['Total2'] = {'Total2':int(rows[1]) * 5}

outFile.writeheader()

for stuff in file1Dict:
    file1Dict[stuff]['Difference'] = str(int(int(file1Dict[stuff]['Total2']) / int(file1Dict[stuff]['Total1'])) * 100) + '\%'
    outFile.writerow(file1Dict[stuff])

Just a quick put together of what you described, without non standard modules.