Pandas: What are the cases when count returned by DataFrame describe is a floating point
Question:
When describing my Pandas dataframe: I get the following result:
Mains_1_Power Mains_2_Power
count 17.000000 17.000000
mean 57.063528 200.428607
std 67.605151 69.364919
min 11.015203 135.492259
25% 31.850638 161.546607
50% 35.871114 183.986024
75% 56.419915 210.772911
max 312.787603 446.077603
I do not understand the cases where count should be a float, do we have half records?
If the count is always an integer, is it just the representation, which is floating point?
It can be confusing to put up results where count is not an integer. Is there a workaround?
Answers:
Count will always be an integer with type float64
(rather than type int
).
This is purely for presentation, because DataFrames columns must have the same type (or, at least, there are performance gains in them doing so over object
type).
When describing my Pandas dataframe: I get the following result:
Mains_1_Power Mains_2_Power
count 17.000000 17.000000
mean 57.063528 200.428607
std 67.605151 69.364919
min 11.015203 135.492259
25% 31.850638 161.546607
50% 35.871114 183.986024
75% 56.419915 210.772911
max 312.787603 446.077603
I do not understand the cases where count should be a float, do we have half records?
If the count is always an integer, is it just the representation, which is floating point?
It can be confusing to put up results where count is not an integer. Is there a workaround?
Count will always be an integer with type float64
(rather than type int
).
This is purely for presentation, because DataFrames columns must have the same type (or, at least, there are performance gains in them doing so over object
type).