On Sun, Nov 1, 2009 at 10:20 PM, David Cournapeau <da...@ar.media.kyoto-u.ac.jp> wrote: > Thomas Robitaille wrote: >> Hi, >> >> I'm trying to generate random 64-bit integer values for integers and >> floats using Numpy, within the entire range of valid values for that >> type. To generate random 32-bit floats, I can use: >> >> np.random.uniform(low=np.finfo(np.float32).min,high=np.finfo >> (np.float32).max,size=10) >> >> which gives for example >> >> array([ 1.47351436e+37, 9.93620693e+37, 2.22893053e+38, >> -3.33828977e+38, 1.08247781e+37, -8.37481260e+37, >> 2.64176554e+38, -2.72207226e+37, 2.54790459e+38, >> -2.47883866e+38]) >> >> but if I try and use this for 64-bit numbers, i.e. >> >> np.random.uniform(low=np.finfo(np.float64).min,high=np.finfo >> (np.float64).max,size=10) >> >> I get >> >> array([ Inf, Inf, Inf, Inf, Inf, Inf, Inf, Inf, Inf, Inf]) >> >> Similarly, for integers, I can successfully generate random 32-bit >> integers: >> >> np.random.random_integers(np.iinfo(np.int32).min,high=np.iinfo >> (np.int32).max,size=10) >> >> which gives >> >> array([-1506183689, 662982379, -1616890435, -1519456789, 1489753527, >> -604311122, 2034533014, 449680073, -444302414, >> -1924170329]) >> >> but am unsuccessful for 64-bit integers, i.e. >> >> np.random.random_integers(np.iinfo(np.int64).min,high=np.iinfo >> (np.int64).max,size=10) >> >> which produces the following error: >> >> OverflowError: long int too large to convert to int >> >> Is this expected behavior, or are these bugs? >> > > I think those are bugs, but it may be difficult to fix. > > You can check that if you restrict a tiny bit your interval, you get > better result: > > import numpy as np > # max/min for double precision is ~ 1.8e308 > low, high = -1e308, 1e308 > np.random.uniformat(low, high, 100) # bunch of inf > low, high = -1e307, 1e307 > np.random.uniformat(low, high, 100) # much more reasonable > > It may be that you are pushing the limits of the random generator. Your > min and max may be border cases: if you use the min/max representable > numbers, and the random generator needs to do any addition of a positive > number, you will 'overflow' your float number (Robert will have a better > answer to this). The problem is that it may be difficult to detect this > in advance. > > David > _______________________________________________ > NumPy-Discussion mailing list > NumPy-Discussion@scipy.org > http://mail.scipy.org/mailman/listinfo/numpy-discussion >
array([ Inf, Inf, Inf, Inf, Inf, Inf, Inf, Inf, Inf, Inf]) might actually be the right answer if you want a uniform distribution on the real line. I never realized how many numbers are out there when I saw that most numbers in the example are e+37 or e+38. Josef _______________________________________________ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion
