Python Pandas - Converting to timestamp

---

---

As we know that Pandas Timestamp represents specific points in time, like Python datetime.datetime. In Pandas, you can convert various objects representing date and time into timestamp using the to_datetime() function, which helps manage and analyze time-series data with ease.

In this tutorial, we will learn about the methods for converting various date formats to Pandas timestamps, converting timestamps back to epoch format, and handling time zones.

Converting Epoch Timestamps to Timestamps

Pandas supports converting integer or float epoch times (i.e., time since January 1, 1970 UTC) into timestamps using the pd.to_datetime(), specifying the unit like "s" for seconds, "ms" for milliseconds, etc.

Example

The following example converts an integer list representing seconds since the epoch into timestamp using the to_datatime() method.

import pandas as pd

# Create a list of integer epoch times	
epoch_seconds = [1439720105, 143806505]

# Convert seconds since epoch to Timestamp
print('Converted epoch seconds to timestamp:')
print(pd.to_datetime(epoch_seconds, unit="s"))

Following is the output of the above code −

Converted epoch seconds to timestamp:
DatetimeIndex(['2015-08-16 10:15:05', '1974-07-23 10:15:05'], dtype='datetime64[ns]', freq=None)

Converting Timestamps Back to Epoch

To convert timestamps back to epoch format can be done by subtracting the epoch (i.e., time since January 1, 1970 UTC) and then apply the floor division by the "unit" (1 second). This is helpful for transforming dates back to raw time data for calculations or storage purposes.

Example

The following example converts list of Pandas timestamps back to epoch format.

import pandas as pd

# Create timestamp 
timestamps = pd.date_range("2024-11-04 18:15:05", periods=4, freq="D")

# Display the input timestamps 
print("Input Timestamps:")
print(timestamps)

# Convert a timestamp to epoch time in seconds
epoch_time = (timestamps - pd.Timestamp("1970-01-01")) // pd.Timedelta("1s")

print("\nOutput Timestamp to epoch time in seconds:")
print(epoch_time)

Following is the output of the above code −

Input Timestamps:
DatetimeIndex(['2024-11-04 18:15:05', '2024-11-05 18:15:05',
               '2024-11-06 18:15:05', '2024-11-07 18:15:05'],
              dtype='datetime64[ns]', freq='D')

Output Timestamp to epoch time in seconds:
Index([1730744105, 1730830505, 1730916905, 1731003305], dtype='int64')

Converting Timestamp to Python Datetime

To convert a Pandas Timestamp object to a native Python datetime object can be done by using the to_pydatetime() method.

Example

Following example demonstrates converting the Pandas Timestamp object to a native Python datetime object using the to_pydatetime() method.

import pandas as pd

# Create timestamp 
timestamps = pd.date_range("2024-11-04 18:15:05", periods=4, freq="D")

# Display the input timestamps 
print("Input Timestamps:")
print(timestamps)

# convert timestamp to native Python datetime object
print("\nConverted Python datetime objects:")
print(timestamps.to_pydatetime())

Following is the output of the above code −

Input Timestamps:
DatetimeIndex(['2024-11-04 18:15:05', '2024-11-05 18:15:05',
               '2024-11-06 18:15:05', '2024-11-07 18:15:05'],
              dtype='datetime64[ns]', freq='D')

Converted Python datetime objects:
[datetime.datetime(2024, 11, 4, 18, 15, 5)
 datetime.datetime(2024, 11, 5, 18, 15, 5)
 datetime.datetime(2024, 11, 6, 18, 15, 5)
 datetime.datetime(2024, 11, 7, 18, 15, 5)]

Converting PeriodIndex to Timestamp

A PeriodIndex object holds ordinal values indicating regular periods in time. Which is nothing but a span of time with a specific frequency. Converting a PeriodIndex object to timestamp can be possible by using the to_timestamp() method.

Example

The following example demonstrates converting the PeriodIndex object to DatetimeIndex (timestamp) using the to_timestamp() method.

import pandas as pd

# Create a PeriodIndex object
periodIndex = pd.PeriodIndex(['2021-09-25 07:30:35', '2019-10-30 04:15:45',
'2020-07-15 02:55:15', '2022-06-25 09:40:55'], freq="Y")

# Display PeriodIndex object
print("PeriodIndex...\n", periodIndex)

# Display PeriodIndex frequency
print("\nPeriodIndex frequency object...\n", periodIndex.freq)

# Display PeriodIndex frequency as string
print("\nPeriodIndex frequency object as a string...\n", periodIndex.freqstr)

# Convert PeriodIndex to timestamp
print("\nPeriodIndex object to timestamp...\n", periodIndex.to_timestamp())

This will produce the following output −

PeriodIndex...
 PeriodIndex(['2021', '2019', '2020', '2022'], dtype='period[A-DEC]')

PeriodIndex frequency object...
 <YearEnd: month=12>

PeriodIndex frequency object as a string...
 A-DEC

PeriodIndex object to timestamp...
 DatetimeIndex(['2021-01-01', '2019-01-01', '2020-01-01', '2022-01-01'], dtype='datetime64[ns]', freq=None)

Convert DataFrame Column to Timestamp

To create Timestamps from multiple DataFrame columns (e.g., year, month, day), you can use pd.to_datetime() method on the selected columns of a DataFrame.

Example

The following example shows the creation of Timestamps from multiple DataFrame columns using the pd.to_datetime() method.

import pandas as pd

# Create the DataFrame with multiple columns
df = pd.DataFrame({"year": [2024, 2023], "month": [2, 3], "day": [4, 5]})

# display the input DataFrame
print("Input DataFrame:")
print(df)

# Convert multiple DataFrame columns to timestamps 
print("\nConverted Timestamp:")
print(pd.to_datetime(df))

Following is the output of the above code −

Input DataFrame:
   year  month  day
0  2024      2    4
1  2023      3    5

Converted Timestamp:
0   2024-02-04
1   2023-03-05
dtype: datetime64[ns]