Importing CSV after initially exporting datetime output (without using pandas)

Question

I use the following to write CSV lines in a loop while running sensor instrumentation, and I want to read them back in for post-processing without pandas. I would like to use the imported data to plot, and do basic arithmetic on the time stamp column, e.g. calculate time separation between events captured by a sensor.

from datetime import datetime

date_time = datetime.now()

filename.write(str(date_time)+','+str(SensorData1)+','+str(SensorData2))

Each line of the CSV output looks like this:

2025-06-01 19:05:24.384927,8.001,4.999

The link suggested here: How do I read and write CSV files? does not automatically deal with datetime.now() formats.

As written to CSV, datetime.now() is a quasi-numerical value that can be used for plotting by matplotlib. So, I am not interested in importing the datetime.now() data from CSV as a string, or some other object, but as one where I can do basic math and use numpy to calculate time-increments between each sensor reading value.

Answer 1

Just use strptime with the format string for now():

>>> from datetime import datetime
>>> fmt='%Y-%m-%d %H:%M:%S.%f'
>>> s=str(datetime.now())
>>> datetime.strptime(s, fmt)
datetime.datetime(2025, 7, 28, 6, 40, 12, 761335)
>>> s==str(datetime.strptime(s, fmt))
True

(Note: The default format for now() is close to ISO 8601 but lacks the required timezone information. It is local time on the machine with the same fields as ISO 8601 output without the timezone field at the end.)

Better still, use the same format string and strftime for the output of now to then parse the same string in your CSV file:

>>> s2=datetime.now().strftime(fmt)
>>> s2==str(datetime.strptime(s2, fmt))
True

Yet another way is to forget using a formatted time string altogether and to use a POSIX time stamp from system epoch time:

>>> datetime.now().timestamp()
1753701765.549574

That just becomes a float representation in your CSV and can then be used for arithmetic directly.