How do you create variable code in Python?

Question

Imagine you have a structured array, for example like this:

import numpy as np


a = np.array(
    [tuple([np.random.randint(100) for _ in range(3)]) for _ in range(100)], 
    dtype=[('var1', 'i4'), ('var2', 'i4'), ('var3', 'i4')]
)

Now I only want to access a specific subset / slice of this array. For example like this:

interval = (10, 30)
b = a[
    (a['var1'] >= interval[0]) & (a['var1'] <= interval[1])
]

So far so good. But what if I have a variable number of intervals corressponding to different variables? For example like this:

intervals = [('var1', 10, 30), ('var2', 20, 50)]

I cannot hardcode it because the amount of intervals changes while the program is running. but what I would like is something like this for an arbitrary number of intervals:

b = a[
    ((a[intervals[0][0]] >= intervals[0][1]) & (a[intervals[0][0]] <= intervals[0][2])) |
    ((a[intervals[1][0]] >= intervals[1][1]) & (a[intervals[1][0]] <= intervals[1][2]))
]

The only idea that I have had so far is using a for loop to go over the intervals and create a string that can then be excecuted using eval, but I don't really like this. Is there a better solution?

string = 'a[((a[intervals[0][0]] >= intervals[0][1]) & (a[intervals[0][0]] ' \
         '<= intervals[0][2]))'
for i in range(len(intervals[1:])):
    string += f' | \n((a[intervals[{i+1}][0]] >= intervals[{i+1}][1]) & ' \
              f'(a[intervals[{i+1}][0]] <= intervals[{i+1}][2]))'
string += ']'

b = eval(string)

MisterMiyagi · Accepted Answer · 2020-01-23 13:19:38Z

2

You don't need to generate code for this problem. Use functools.reduce to apply the | operator on an unknown series of masks.

import operator
from functools import reduce

mask = reduce(
    operator.__or__,  # the `|` operator
    (
        ((a[var] >= lower) & (a[var] <= upper))
        for var, lower, upper in intervals  # arbitrary number of intervals
    )
)

b = a[mask]

answered Jan 23, 2020 at 13:19

MisterMiyagi

53.4k14 gold badges131 silver badges138 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Tarifazo · Accepted Answer · 2020-01-23 13:20:00Z

1

I'd put the conditions in a list of tuples, and bind them using a function and map, i.e.:

def fmask(tup):
  ix, low, up = tup
  return (a[ix] > low) & (a[ix] < up)

conds = [('var1', 10, 80), ('var2', 30, 40)]

a[np.logical_and(*map(fmask))]

answered Jan 23, 2020 at 13:20

Tarifazo

4,3631 gold badge11 silver badges24 bronze badges

Collectives™ on Stack Overflow

How do you create variable code in Python?

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related