How to Convert Datetime to Date in Pandas

Often you may want to convert a datetime to a date in pandas. Fortunately this is easy to do using the .dt.date function, which takes on the following syntax:

df['date_column'] = pd.to_datetime(df['datetime_column']).dt.date

Example: Datetime to Date in Pandas

For example, suppose we have the following pandas DataFrame:

import pandas as pd

#create pandas DataFrame with two columns
df = pd.DataFrame({'sales': [4, 11],
                   'time': ['2020-01-15 20:02:58', '2020-01-18 14:43:24']})

#view DataFrame 
print(df)

sales	time
0	4	2020-01-15 20:02:58
1	11	2020-01-18 14:43:24

To convert the ‘time’ column to just a date, we can use the following syntax:

#convert datetime column to just date
df['time'] = pd.to_datetime(df['time']).dt.date

#view DataFrame
print(df)

	sales	time
0	4	2020-01-15
1	11	2020-01-18

Now the ‘time’ column just displays the date without the time.

Using Normalize() for datetime64 Dtypes

You should note that the code above will return an object dtype:

#find dtype of each column in DataFrame
df.dtypes

sales     int64
time     object
dtype:   object

If you instead want datetime64 then you can normalize() the time component, which will keep the dtype as datetime64 but it will only display the date:

#convert datetime column to just date
df['time'] = pd.to_datetime(df['time']).dt.normalize()

#view DataFrame
print(df)

	sales	time
0	4	2020-01-15
1	11	2020-01-18

#find dtype of each column in DataFrame
df.dtypes

sales             int64
time     datetime64[ns]
dtype: object

Once again only the date is displayed, but the ‘time’ column is a datetime64 dtype.

Additional Resources

How to Convert Columns to DateTime in Pandas
How to Convert Strings to Float in Pandas