This script compares values of each attribute with a "central value" if the values deviate within a certain threshold from the central value, set the value to the central value.

zeroSmallValues_v001_02.py

'''
written by Patrick Woo ([email protected])
20230309
version 001_02

change log
v001_02
- now the tool reports number of attributes and number of objects changed

This script compares values of each attribute with a "central value"
if the values deviate within a certain threshold from the central value, 
set the value to the central value.

This is good for "fixing" attribute values that show up in channel boxes as -0
when the actual value is a very small number like 0.00005452.
In these cases we may want to just set them to 0 rightaway

eg, 
    when ctrVal is 0, threhsold is 0.009, these values will result from various incoming values
        0.001 => 0
        0.00042 => 0
        0.009 => 0.009
        0.0089 => 0
        0.0091 => 0.0091
        -0.00045 => 0
        -0.009 => -0.009
        -0.0089 => 0
        -0.0091 => -0.0091
'''

from pprint import pprint as pp
import pymel.core as pm
# pm.getAttr('.rx')

def zeroSmallValues(attrList, ctrVal=0, thresh=0.0009):
    '''
    zero small values
    rounds out small values. read the main doc string above
        
    attrList - <list> a list of string attributes that will be processed. examples:
                ['tx', 'ty', 'tz']
                ['translateX', 'translateY', 'translateZ']
                ['object001.tx', 'object001.rotateY']
    ctrVal - <float> the centre value which small fluctuations of value within the threshold range will be snapped to.
    thresh - <float> the threshold value which if the values are lesser than this threshold will be ignored and set to the centre value
    '''
    
    print(attrList)
    changedAttrCounter = 0
    changedObjs = [] # contains the names of the changed objects
    for thisAttr in attrList:
        valDiff = abs(ctrVal - thisAttr.get())
        if valDiff < thresh:
            thisAttr.set(ctrVal)
            print('{} set to {}'.format(thisAttr, ctrVal))
            changedAttrCounter += 1
            # add the name of the changed object to changedObjs
            if not thisAttr.node() in changedObjs:
                changedObjs.append(thisAttr.node())
        # end if valDiff
    # end for thisAttr
    print('')
    print('{} attributes changed in {} objects'.format(changedAttrCounter, len(changedObjs)))
    return

# ----
# main
# ----
# get all selected objects. children in hierarchies need to be explicitly selected
# (not just by selecting the parents)
objList = pm.ls(sl=True)
targetAttrList = ['tx', 'ty', 'tz', 'rx', 'ry', 'rz']
attrList = [pm.PyNode(x).attr(y) for x in objList for y in targetAttrList]

# pp(attrList) # print out for dignostics

# call the function
zeroSmallValues(attrList, ctrVal=0, thresh=0.0009)