有条件地从python列表中提取数字

Question

I have a list of numbers like

20
40
45
60
80

That I want to be able to say, for example the average distance between numbers < 50 is 12.5.

import numpy as np
from sys import argv
script, pos_file, output = argv
positions = []
with open(pos_file) as f:
    for x in f:
        assert x.strip().split()
        positions.append(x)

position_list= []

for x in positions:
    if x < 50:
        position_list.append(x)

print np.mean[position_list]

this doesn't work - I think because when I print the positions list I get 20,40,45,60,80 - so I think it is not treating the numbers as individual numbers hence it cannot test if x < 50. What am I doing wrong?

EDIT: looks like the data is rather made of lines like:

467,1977,3751,4013,5752,6406,6446,7362,7585,8285,8624,8741,‌​9143,9304,11879,1319‌​7,13460,14401,14785,‌​15117,22264,23714,24‌​294,24534,26053,2695‌​9,27714,29462,35342,‌​36538,36612,37031,39‌​093,42281,42967,4394‌​5

Answer 1

There are several things wrong with your code:

• you do not convert them to an int or float ;
• you use np.mean[..] instead of np.mean(..) and np.mean is not scriptable.

The solution is:

import numpy as np
from sys import argv
script, pos_file, output = argv
positions = []
with open(pos_file) as f:
    for x in f:
        assert x.strip().split()
        positions.append(int(x))

position_list= [x for x in positions if x < 50]

print np.mean(position_list)

EDIT

Based on your comments however, it looks like you feed a comma separated list:

import numpy as np
from sys import argv
script, pos_file, output = argv
positions = []
with open(pos_file) as f:
    for x in f:
        positions += (int(i) for i in x.strip().split())

position_list= [x for x in positions if x < 50]

print np.mean(position_list)

Or:

import numpy as np
from sys import argv
script, pos_file, output = argv
positions = []
with open(pos_file) as f:
    for x in f:
        for i in x.strip().split():
            positions.append(int(i))

position_list= [x for x in positions if x < 50]

print np.mean(position_list)

You can also, as @Jean-FrançoisFabre says, use the sum and divide by the number of items, so:

from sys import argv
script, pos_file, output = argv
positions = []
with open(pos_file) as f:
    for x in f:
        for i in x.strip().split():
            positions.append(int(i))

position_list= [x for x in positions if x < 50]

print sum(position_list)/len(position_list)

In that case you do not have to import numpy .

Answer 2

There are a few errors in your code, other answers pointed that out, but I feel I should rewrite it for you in a much cleaner way:

with open(pos_file) as f:
    positions = [int(x) for line in f for x in line.strip().split(',') if int(x) < 50]

print(sum(positions)/len(positions))

• you don't need numpy to compute a mean , this isn't rocket science
• the assert statement is useless. If a line is empty, split() returns an empty list, not a problem for the list comprehension.
• the added double loop allows to read several integers located on the same line
• no memory wasted storing numbers when you only want to keep the lowest ones
• took advantage of your feedback on one of the answers to figure out that the list is comma separated. Now I realize that the csv module could have been used.

so csv solution:

import csv
with open(pos_file) as f:
    cr = csv.reader(f)
    positions = [int(x) for row in cr for x in row if int(x) < 50]

print(sum(positions)/len(positions))