Float formatting

[Axel_Erfurt]

maybe astype(float)

df3['T'] = np.nan_to_num(df3['T']).astype(float)

Output:
0    11.00
1    22.00
2    11.23
3    20.03
Name: T, dtype: float64