Pandas None series has 0

Question

I want to generate a pandas series with null values, with int type. I use this:

import pandas as pd
pd.Series(None, index=[1, 2, 3], dtype=int)

And it returns

1    0
2    0
3    0
dtype: int64

What can I do for it to return a series whose values are null?

jezrael · Answer 1 · 2021-12-21T08:14:18.703

3

s = pd.Series(None, index=[1, 2, 3], dtype='Int64')
print (s)
1    <NA>
2    <NA>
3    <NA>
dtype: Int64

s = pd.Series(np.nan, index=[1, 2, 3], dtype=int)
print (s)
1   NaN
2   NaN
3   NaN
dtype: float64 <- casting to float

edited Dec 21 '21 at 08:14

answered Dec 21 '21 at 08:04

jezrael

My main issue is that I'm trying to create series from the type of a variable: ```import pandas as pd; a = 2; type_out = type(a); pd.Series(None, index=[1, 2, 3], dtype=type_out)``` – David Masip Dec 21 '21 at 08:10
And I don't know how can I go from int to int64, float to float64, and all the other combinations – David Masip Dec 21 '21 at 08:10
@DavidMasip - It is not possible `type=int` and `None` or `NaN` – jezrael Dec 21 '21 at 08:11
@DavidMasip - only possible some default value, like `0` instead `None` – jezrael Dec 21 '21 at 08:12
And why is this not good? ```pd.Series(np.nan, index=[1, 2, 3], dtype=int)``` – David Masip Dec 21 '21 at 08:12
@DavidMasip - because it failed returning `int`. – jezrael Dec 21 '21 at 08:12
1

@DavidMasip - it return floats – jezrael Dec 21 '21 at 08:13
This post is relevant: https://stackoverflow.com/questions/12708807/numpy-integer-nan – nikeros Dec 21 '21 at 08:29

1 Answers1