Finding multiple of a decimal number in a floating point list

[Alejandro]

Hi:

I need to find the multiples of a decimal number in a floating point
list. For instance, if a have the list [0,0.01,0.02,... 1], I want the
multiples of 0.2: [0, 0.2,0.4,0.6,0.8 ,1].

With integers this problem is easy, just test for (i%n == 0), where i
is the number I am testing, and n is the multiple. Given the finite
resolution of floating point numbers, this is more complicated for
float.

I came with this solution:

from numpy import arange
from math import modf, fabs

float_range = arange(0, 1, 0.01)
for f in float_range:
m = modf(f / 0.2)[0]
if m<1e-13 or fabs(1-m)<1e-13:
print f
# Do something else

This code works, however, I found it a little ugly. Is there a better
way to do the same?

Alejandro.

[Dan Lenski]

On Mon, 18 Aug 2008 10:52:45 -0700, Alejandro wrote:

Hi:

I need to find the multiples of a decimal number in a floating point
list. For instance, if a have the list [0,0.01,0.02,... 1], I want the
multiples of 0.2: [0, 0.2,0.4,0.6,0.8 ,1].

With integers this problem is easy, just test for (i%n == 0), where i is
the number I am testing, and n is the multiple. Given the finite
resolution of floating point numbers, this is more complicated for
float.

I came with this solution:

from numpy import arange
from math import modf, fabs

float_range = arange(0, 1, 0.01)
for f in float_range:
m = modf(f / 0.2)[0]
if m<1e-13 or fabs(1-m)<1e-13:
print f
# Do something else

This code works, however, I found it a little ugly. Is there a better
way to do the same?

Alejandro.

Hi Alejandro, you can do the same thing more efficiently (both in terms
of lines of code and execution speed) by doing the whole array at once:

from numpy import arange, absolute
from math import modf, fabs

float_range = arange(0, 1, 0.01)
multiples = absolute(float_ range % 0.2)<1e-13
# now multiples is a boolean array
print float_range[multiples]

HTH,
Dan

[Dan Lenski]

On Mon, 18 Aug 2008 18:04:53 +0000, Dan Lenski wrote:

On Mon, 18 Aug 2008 10:52:45 -0700, Alejandro wrote:

>Hi:

I need to find the multiples of a decimal number in a floating point
list. For instance, if a have the list [0,0.01,0.02,... 1], I want the
multiples of 0.2: [0, 0.2,0.4,0.6,0.8 ,1].

With integers this problem is easy, just test for (i%n == 0), where i
is the number I am testing, and n is the multiple. Given the finite
resolution of floating point numbers, this is more complicated for
float.

I came with this solution:

from numpy import arange
from math import modf, fabs

float_range = arange(0, 1, 0.01)
for f in float_range:
m = modf(f / 0.2)[0]
if m<1e-13 or fabs(1-m)<1e-13:
print f
# Do something else

This code works, however, I found it a little ugly. Is there a better
way to do the same?

Alejandro.

Hi Alejandro, you can do the same thing more efficiently (both in terms
of lines of code and execution speed) by doing the whole array at once:

from numpy import arange, absolute
from math import modf, fabs

float_range = arange(0, 1, 0.01)
multiples = absolute(float_ range % 0.2)<1e-13 # now multiples is a
boolean array
print float_range[multiples]

HTH,
Dan

Oh, I forgot that you have to check the other case of it being slightly
less than a multiple too!

from numpy import arange, absolute
from math import modf, fabs

float_range = arange(0, 1, 0.01)
mod = absolute(float_ range % 0.2)
multiples = (mod < 1e-13) + (mod 0.2-1e-13)
print float_range[multiples]

That should do the trick!

Dan