分配给循环值

Question

This is really two questions.

I have a list of age intervals. For each interval there is a corresponding value. The intervals and values are organized in a list of tuples age_value_intervals (see the comments in the code).

I also have a separate list of distinct ages, ages , for which I would like to know the value.

The code below is an attempt at mapping the value to the given age.

Now to the questions,

1. In order to assign a value to the value_map I iterate over both ages and value_map using zip . I then try to assign to value . This doesn't work. Why?

2. I doubt that the method I use is very efficient (if it had worked). Is there a better way to achieve this mapping?

---

import numpy as np

# List of tuples defining and age interval and the corresponing value for
# that interval. For instance (20, 30, 10) indicates that the age interval from
# 20 to 30 has the value 10
age_value_intervals = [(20, 30, 10),
                       (30, 35, 5),
                       (35, 42, 50),
                       (50, 56, 40),
                       (56, 60, 30)]

# The ages for which I would like to know the value
ages = [25, 30, 35, 40, 45, 50]

# Empty array used to stor the values for the corresponding age
value_map = np.empty(len(ages))
# I want the value to be nan if there is no known value
value_map[:] = np.nan

# Iterate over the ages I want to know the value for
for age, value in zip(ages, value_map):
    # Check if the age is in an interval for which the value is known
    for from_age, to_age, actual_value in age_value_intervals:
        if age >= from_age and age < to_age:
            # Assign the value to the value_map
            # This is were it falls apart (I guess...)
            value = actual_value
            # Move on to the next age since we got a match
            break

#Expected output
value_map = [10, 5, 50, 50, nan, 40]

Answer 1

I recommend you use numpy.digitize together with dict for this. You can manually account for instances when a value cannot be mapped to a range.

import numpy as np

age_value_intervals = [(20, 30, 10),
                       (30, 35, 5),
                       (35, 42, 50),
                       (50, 56, 40),
                       (56, 60, 30)]

ages = np.array([25, 30, 35, 40, 45, 50])

bins = np.array([x[0] for x in age_value_intervals])
mapper = dict(enumerate([x[2] for x in age_value_intervals], 1))    

res = np.array([mapper[x] for x in np.digitize(ages, bins)], dtype=float)

for idx in range(len(ages)):
    if not any(i <= ages[idx] <= j for i, j, k in age_value_intervals):
        res[idx] = np.nan

Result:

array([ 10.,   5.,  50.,  50.,  nan,  40.])

Answer 2

First, as noted on comments, if you try to assign to variable you are currently changing inside loop, the value simply gets lost.

Secondly most of the mappings are redundant.

Something like this can probably still be improved but should work:

result=[] 
for check_age in ages:
    for from_age, to_age, value in age_value_intervals:
        if check_age in range(from_age, to_age):
            result+=[value]

print result

Note, if you need some result added also when the age is not in the interval, there needs to be additional code.