Nicely representing a floating-point number in python [duplicate]

Answer 1

Here is a snippet that formats a value according to the given error bars.

from math import floor, log10, round

def sigfig3(v, errplus, errmin):
    i = int(floor(-log10(max(errplus,errmin)) + 2))
    if i > 0:
        fmt = "%%.%df" % (i)
        return "{%s}^{%s}_{%s}" % (fmt % v,fmt % errplus, fmt % errmin)
    else:
        return "{%d}^{%d}_{%d}" % (round(v, i),round(errplus, i), numpy.round(i))

Examples:

5268685 (+1463262,-2401422) becomes 5300000 (+1500000,-2400000)
0.84312 +- 0.173124 becomes 0.84 +- 0.17