I Have series of float numbers - [0.1, 0.01, 0.0001, 0.000123]. How can i get "longest" float (not a minimum!) - 0.000123?

CodePudding user response：

Convert the float numbers to string and then use len with pandas.Series.argmax :

import pandas as pd

s = pd.Series([0.1, 0.01, 0.0001, 0.000123])

float_length = s.astype(str).map(len)
out = s.loc[float_length.argmax()]

# Output :

print(out)

0.000123

CodePudding user response：

try:

import numpy as np

# take a list/numpy array of floats and return the longest float
def get_longest(x):
    idx = sorted([(n, len(str(i).split('.')[-1])) for n, i in enumerate(x)], key=lambda x: x[1])[-1][0]
    
    return x[idx]

a = [0.000123, 0.1, 0.01, 0.0001]

print(get_longest(a))
# 0.000123

print(get_longest(np.array(a)))
# 0.000123

b = [0.1, 0.01, 0.0001, 0.000123]

print(get_longest(b))
# 0.000123

if you have scientific notation (float with very long Decimal parts) in your list/np array:

def get_longest(x):
    x = [str(np.format_float_positional(i, trim='-')) for i in x]
    idx = sorted([(n, len(i.split('.')[-1])) for n, i in enumerate(x)], key=lambda x: x[1])[-1][0]
    
    return x[idx]

c = [9e-09, 6.5e-20, 0.1, 0.01] # --> [0.000000009, 0.000000000000000000065, 0.1, 0.01]

print(get_longest(np.array(c)))
# 0.000000000000000000065

print(get_longest(c))
# 0.000000000000000000065