How to Convert Between Python's `datetime`, Pandas' `Timestamp`, and NumPy's `datetime64`?

Converting Between datetime, Timestamp, and datetime64

This article explores the conversion process between datetime, Timestamp, and datetime64 objects in Python.

Numpy's datetime64 stores dates and times as 64-bit integers representing nanoseconds since a specified epoch. On the other hand, datetime and Timestamp represent dates and times using Python's datetime class and pandas' Timestamp class, respectively.

To convert from datetime64 to datetime or Timestamp, we can utilize the pd.Timestamp constructor. The following conversion diagram provides a comprehensive overview:

[Image of Conversions between time representations]

Conversions from datetime64

dt64 = np.datetime64('2012-05-01T01:00:00.000000+0100')

# Converting to datetime
dt = pd.Timestamp(dt64).to_datetime()

# Converting to Timestamp
ts = pd.Timestamp(dt64)

Conversions from datetime

dt = datetime.datetime(2012, 5, 1, 1, 0)

# Converting to datetime64
dt64 = pd.Timestamp(dt).to_datetime64()

# Converting to Timestamp
ts = pd.Timestamp(dt)

Conversions from Timestamp

ts = pd.Timestamp('2012-05-01 01:00:00')

# Converting to datetime
dt = ts.to_datetime()

# Converting to datetime64
dt64 = ts.to_datetime64()

Note:

It is important to pay attention to the possible timezone offset when dealing with datetime64 objects.

The above is the detailed content of How to Convert Between Python's `datetime`, Pandas' `Timestamp`, and NumPy's `datetime64`?. For more information, please follow other related articles on the PHP Chinese website!