如何将字典列表值转换为整数？

Question

I'm working in Python 3.5. I've uploaded a CSV file and made it into a dictionary. However, the list of multiple values for each key is a string and not an integer. How can I convert the values for each key into an integer?

Furthermore, is there a way for future CSV importation to automatically make dictionary value lists into integers?

So far this is what I have:

import csv
reader = csv.reader(open('filename.csv'))
dictname = {}
for row in reader:
    key = row[0]
    if key in dictname:
        pass
    dictname[key] = row[1:]

print dictname

Answer 1

if row is the list containing the integers under string format :

dictname[key] = [int(elt) for elt in row[1:] if elt.isdigit()]

should do the trick

Answer 2

You can use pandas and specify converter functions. In fact you may not even need to do that as it is intelligently parsing CSV files.

import pandas as pd

df = pd.read_csv('filename.csv')

If you need converter functions:

df = pd.read_csv('filename.csv',converters={'yourintegercolumn':int})

Answer 3

I used a function to first check if the value is a string/unicode. If so, it then tries to convert it to a float, eg "1,234.45" -> 1234.45. If this fails or the value is not a string/float, the function returns it unchanged.

This function is then used in a list comprehension to population the dictionary.

Note that the if key in dictname: pass block doesn't do anything. If there are duplicate keys in your data, you have three options:

1) Overwrite the existing key's data with the new row that has the identical key value (this is what is currently happening).

2) Only use the first occurrence of the key row. In this case, change pass to continue .

3) Try to aggregate the data. This is more complicated and beyond the scope of your original question, so I will leave it to you to figure out or post a new question covering this scope.

def convert_to_numeric(value):
    if isinstance(i, (str, unicode)):
        try:
            result = float(value)
        except:
            pass  # Returns result on next line.
    return result

for row in reader:
    key = row[0]
    if key in dictname:
        pass  # This doesn't do anything. Use `continue` to avoid overwriting.
    dictname[key] = [convert_to_numeric(i) for i in row[1:]]