(Jul-20-2019, 08:48 AM)scidam Wrote: Ability to use NaN values with columns of integer type is latest pandas feature (and still experimental).
As of Pandas 0.24.x you can do this, e.g.
import pandas as pd import numpy as np diSales= { 2016:{'qtr1':34500,'qtr2':56000,'qtr3':47000,'qtr4':49000}, 2017:{'qtr1':44900,'qtr2':46100,'qtr3':57000,'qtr4':59000}, 2018:{'qtr1':54500,'qtr2':51000,'qtr3':57000,'qtr4':58500}, 2019:{'qtr1':61000} } df = pd.DataFrame(diSales, dtype=pd.Int32Dtype()) df.mad(axis=0)df.info()
returns integer dtype, butdf.mad
still returns float.
Output:2016 6062.5 2017 6250.0 2018 2500.0 2019 0.0 dtype: float64
Still showing dtyle : float64 in output after using your code which have line: df = pd.DataFrame(diSales, dtype=pd.Int32Dtype()).why ? What is use of this line ?