How to replace "values with more than one decimal" with "nan" in a given column-CodePudding

I have a data frame in which one of the columns (dtype = float64) has a set of values such as:

129.0
nan
100.0
87.0
40.0
344.992
130.0
101.0
227.0
147.0
190.0
83.0
-144.63542183979368

I wish to replace all the values with more than one decimal places to nan. The values with more than one decimal places, both positive and negative are actually junk values. Only the positive values with single decimal place ending with .0 are genuine.

So, in above case, -144.63542183979368 and 344.992 should be replaced with nan. The modified data frame column should become like this:

129.0
nan
100.0
87.0
40.0
nan
130.0
101.0
227.0
147.0
190.0
83.0
nan

How do I go about doing this?

At the end, after removing junk float values, I would like to change the dtype to integer (which can be done once the improper float values are removed).

CodePudding user response：

Try This

import pandas as pd
import numpy as np

arr = [129.0 ,np.nan ,100.0 ,87.0, 40.0, 344.992, 130.0, 101.0, 227.0, 147.0, 190.0, 83.0, -144.63542183979368]
df = pd.DataFrame(arr, columns=['example'])
print(df)

def convert(row):
    if row:
        conv = row % 1
        if conv > 0:
            return np.nan
        else:
            return row
    else:
        return np.nan

df['example'] = df['example'].apply(convert)

print(df)