convert pandas dataframe datatypes from float64 into int64

Question

I am trying to read CSV file by using python pandas, in the resultant dataframe one column is returned as float64 datatype instead of int64. But I could see most of the values are numbers and some of them are null values in the existing CSV file

df = pd.read_csv(file)

dh.head(3)

Name State  Id
SFO  CA     123.0
JFK  NY     152.0
CHG  IL     NaN
ABC  AZ     NaN

df.dypes

Name Object
State Object
Id float64

I tried convert Id column into Int64 to upload data into oracle table

df['Id'] = df['Id'].astype('int64')

Error : Cannot convert NA to integer

Is there any approach to convert Id column into int64 ? I appreciate your response.

Please provide a minimal reproducible example, as well as the entire error message. — AMC
– AMC, Commented Mar 31, 2020 at 1:03
Does this answer your question? Convert Pandas column containing NaNs to dtype `int` — AMC
– AMC, Commented Mar 31, 2020 at 1:04

Chicodelarose · Accepted Answer · 2020-03-31 02:17:13Z

1

In Python 3.7.6 and pandas 1.0.3 you can do:

df['Id'] = df['Id'].astype(pd.Int64Dtype())

print(df.dtypes)
print(df)

Output:

Name     object
State    object
Id        Int64

State    Id
0  SFO    CA   123
1  JFK    NY   152
2  CHG    IL  <NA>
3  ABC    AZ  <NA>

edited Mar 31, 2020 at 2:17

answered Mar 31, 2020 at 1:30

Chicodelarose

8827 silver badges24 bronze badges

Sign up to request clarification or add additional context in comments.

5 Comments

N9909 Over a year ago

I have tried, but no luck.Its giving error AttributeError: Module 'pandas' has no attribute 'pd.Int64Dtype()'

Chicodelarose Over a year ago

What is your python and pandas version?

N9909 Over a year ago

python - 3.6 and pandas - 0.19.2

Chicodelarose Over a year ago

@Naresh I added my Python and pandas version to my answer. Good luck!

Chicodelarose Over a year ago

@Naresh You -at least- will have to upgrade to pandas 0.24.2 for this to work pandas.pydata.org/pandas-docs/version/0.24/whatsnew/…

Charts · Accepted Answer · 2024-10-31 04:22:54Z

1

in your code

df['Id'] = df['Id'].astype('int64')

for astype('Int64') in code, try to use cap letter 'I', not small letter 'i', Like

df['Id'] = df['Id'].astype('Int64')

answered Oct 31, 2024 at 4:22

Charts

114 bronze badges

1 Comment

Goran B. Over a year ago

the "int64" literal is valid and doesn't need to be capitalized, best case they can invoke the 'ignore' argument: pandas.pydata.org/docs/reference/api/… ... or as someone suggested use a pd.Int64Dtype() function which handles NaN differently ... keep contributing and best regards.

Collectives™ on Stack Overflow

convert pandas dataframe datatypes from float64 into int64

2 Answers 2

5 Comments

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

5 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related