Generating a Datetime Series in Python

In Excel, you can generate a series of consecutive dates with a simple drag. In Python, you can do something similar with the datetime and timedelta modules. Whether you want intervals of one hour, one day, or one month, you can do it.

It’s actually quite simple. Sometimes your brain just gets stuck and you don’t think of doing it this way. By using timedelta together with a loop, you can generate a series of datetime strings and store them in an array. The following example generates an array of times at 5-second intervals:

import datetime

dt = datetime.datetime(2010, 12, 1,)
end = datetime.datetime(2010, 12, 30)
step = datetime.timedelta(days=1)

result = []

while dt < end:
  result.append(dt.strftime('%Y-%m-%d %H:%M:%S'))
  dt += step

print(result)

Of course, you can also rewrite it with a list comprehension to speed things up, or use a generator depending on the situation.

import datetime as dt

print([dt.date(2017,1,1)+dt.timedelta(days=i) for i in range(30)])

But if we’re talking about the most convenient option, it’s the date_range function provided by Pandas. It wraps the generator for creating a time series into a function for you. It performs well and saves space. Unlike generating values with a loop, which can take a while when the range is large, let’s look at an example:

# Generate 100 items in daily units, starting from 2017/1/1

import pandas as pd

result = pd.date_range('2017/1/1 10:00:00', periods=100, freq='D')

print(result.tolist())
print(result[0:5].to_pydatetime())

for dt in result:
  print(dt.to_pydatetime())

For freq, you can use values other than D. For example, H is hour and M is month. Or refer to the table below: