This function takes two floats, a value and its error/uncertainity (145.7 and 3.4) and returns a string with scientific format (146 ± 3) showing one significant digit error and the correct number of significant digits for the value given the specified error.
| Example: | Output |
|---|---|
| format_error(100, 1) | 100 ±1 |
| format_error(100, 0.001) | 100.000 ±0.001 |
| format_error(3.418, 0.123) | 3.4 ±0.1 |
| format_error(6.3, 0.09) | 6.30 ±0.09 |
| format_error(46288, 1551) | 46000 ±2000 |
| format_error(428.351, 0.27) | 428.4 ±0.3 |
| format_error(0.01683, 0.0058) | 0.017 ±0.006 |
import numpy as np
def format_error(value, error):
# Rounding method used is banker's rounding
digits, exp = np.format_float_scientific(error, 0).split('e') # print as sci. notation, extract digits & exp
decimals = abs(int(exp)) if int(exp) <= 0 else 0
digits = np.around(float(digits))
value = np.around(value, -int(exp)) # adjust significant digits to the ones allowed by the error.
output_string = '{:.'+str(decimals)+'f} ±{:.'+str(decimals)+'f}'
return output_string.format(value, digits * 10 ** int(exp)) #return formatted string