Python/Numpy - Set Values within some Tolerance to the Value

Question

If I have an array and I want to set values 'close' to some value as that value, what is the best way to do this? I'm wondering if their is a numpy function for this. If there is no numpy function, then is the code below the "best" (i.e. quickest/most efficient) way to do this? It works for multi-dimensional arrays as well.

Code:

from numpy import array
tol = 1e-5

# Some array with values close to 0 and 1
t = array([1.0e-10, -1.0e-10, 1.0+1.0e-10, 1.0-1.0e-10, 5.0])
print t[0], t[1], t[2], t[3], t[4]

# Set values within 'tol' of zero to zero
t[abs(t) < tol] = 0.
print t[0], t[1], t[2], t[3], t[4]

# Set values within 'tol' of some value to that value
val = 1.
t[abs(t-val) < tol] = val
print t[0], t[1], t[2], t[3], t[4]

I think that's the best way of 'data cleansing'. I asked a similar question on data cleansing a while ago, and your method is what was recommended. stackoverflow.com/questions/4339273/… — Pete
– Pete, Commented Mar 9, 2011 at 17:31

eat · Accepted Answer · 2011-03-08 19:05:05Z

3

It's not so quite clear what you are trying to achieve, but my interpretation is that around is the solution for your case.

answered Mar 8, 2011 at 19:05

eat

7,5401 gold badge21 silver badges28 bronze badges

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

2 Comments

Scott B Over a year ago

This is probably the closest to what I want. However, it doesn't always do the exact same thing. For example, with my code above, I could set all values in an array within the range of [val-tol,val+tol] to some value, without changing (i.e. rounding) any other values in the array. Say I wanted to change all values in the interval [-3,3] to 0.0 without changing values outside the range of [-3,3], then around wouldn't work. With that said, it does work for just small tolerances and will work for my current purpose.

eat Over a year ago

@Scott B: You may want to work only with those elements fulfilling some conditions. For example L= (-3< t)& (t< 3), t[L]= around(t[L], ...). Thanks

Anycorn · Accepted Answer · 2011-03-08 18:55:52Z

3

http://docs.scipy.org/doc/numpy/reference/generated/numpy.clip.html

answered Mar 8, 2011 at 18:55

Anycorn

51.8k45 gold badges173 silver badges267 bronze badges

1 Comment

Scott B Over a year ago

But this wouldn't round a value of +1.0e-8 to 0.0. It would change a value of -1.0e-8 to 0.0 though (both of these using a clip of something similar to [0,1]). However, I'm glad you pointed out this function as it will be useful for my task as well...

Collectives™ on Stack Overflow

Python/Numpy - Set Values within some Tolerance to the Value

2 Answers 2

2 Comments

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related