How to do random sampling within interval with given interval value?

Question

How can I do random sampling within interval with given interval? For example, I want to do single random sampling between 1 to 10 but with each interval is 0.5. So when I do the sampling, it will give me value for example 5.5 or 2 or 8.5.

I have tried with np.random.random_integers(1,10) but this just give me integers value. Your help is kindly appreciated.

Answer 1

You can write your own function which scales, generates a random integer, and then unscales.

from math import floor, ceil
from random import randint

def random_step(low, high, step):
    low_scaled = ceil(float(low)/step)
    high_scaled = floor(float(high)/step)
    return randint(low_scaled, high_scaled) * step

print random_step(3.0, 8.5, 0.5) # 4.5

The output will always be a multiple of step.

Answer 2

You can use linspace from the numpy library. Here is an example:

>>> import numpy
>>> numpy.linspace(0.5, 10, 20)
array([  0.5,   1. ,   1.5,   2. ,   2.5,   3. ,   3.5,   4. ,   4.5,
         5. ,   5.5,   6. ,   6.5,   7. ,   7.5,   8. ,   8.5,   9. ,
         9.5,  10. ])
>>> range_of_numbers = numpy.linspace(0.5, 10, 20)
>>> random_number = numpy.random.choice(range_of_numbers, size=1)
>>> print random_number
[ 9.5]

The result set of linspace is in an array, but you can easily work with that further if need be. Here is some more documentation on using linspace.

---

EDIT

Another option is to use Pylab's frange library. It's a wrapper around matplotlib, but worth looking into for the solution you are trying to achieve. Here's an example of using frange:

>>> import pylab
>>> pylab.frange(0.5,10,0.5)
array([  0.5,   1. ,   1.5,   2. ,   2.5,   3. ,   3.5,   4. ,   4.5,
         5. ,   5.5,   6. ,   6.5,   7. ,   7.5,   8. ,   8.5,   9. ,
         9.5,  10. ])

Here's some more documentation around frange.

---

Additional Notes

While you could use arange, keep in mind that it leads to floating point errors. The NumPy documentation recommends using linspace to avoid the floating point errors. Also, arange and linspace have different APIs.