Cómo calcular el promedio de pandas pd.Timestamp

Problema

Si tienes un array de objetos pd.Timestamp, no puedes calcular directamente el promedio ya que no pueden ser sumados directamente:

average_timestamp_problem.py

import pandas as pd

# Crear un array de cinco objetos pd.Timestamp fijos
timestamps = [
    pd.Timestamp('2023-01-01 12:00:00'),
    pd.Timestamp('2023-01-02 12:00:00'),
    pd.Timestamp('2023-01-03 12:00:00'),
    pd.Timestamp('2023-01-04 12:00:00'),
    pd.Timestamp('2023-01-05 12:00:00')
]

# FAIL: Esto lanzará un TypeError
average = sum(timestamps) / len(timestamps)

Esto lanzará un TypeError:

error.txt

TypeError                                 Traceback (most recent call last)
Cell In[1], line 13
      4 timestamps = [
      5     pd.Timestamp('2023-01-01 12:00:00'),
      6     pd.Timestamp('2023-01-02 12:00:00'),
   (...)
      9     pd.Timestamp('2023-01-05 12:00:00')
     10 ]
     12 # FAIL: Esto lanzará un TypeError
---> 13 average = sum(timestamps) / len(timestamps)

File timestamps.pyx:483, in pandas._libs.tslibs.timestamps._Timestamp.__radd__()

File timestamps.pyx:465, in pandas._libs.tslibs.timestamps._Timestamp.__add__()

TypeError: Addition/subtraction of integers and integer-arrays with Timestamp is no longer supported.  Instead of adding/subtracting `n`, use `n * obj.freq`

Solución

Puedes sumar/promediar ts.value en lugar de sumar ts directamente, y después del promediado, convertirlo de vuelta a un timestamp:

average_timestamp_solution.py

average = pd.Timestamp(sum(ts.value for ts in timestamps) / len(timestamps))

Ejemplo completo:

average_timestamp_full_example.py

import pandas as pd

# Crear un array de cinco objetos pd.Timestamp fijos
timestamps = [
    pd.Timestamp('2023-01-01 12:00:00'),
    pd.Timestamp('2023-01-02 12:00:00'),
    pd.Timestamp('2023-01-03 12:00:00'),
    pd.Timestamp('2023-01-04 12:00:00'),
    pd.Timestamp('2023-01-05 12:00:00')
]

# Result: Timestamp('2023-01-03 12:00:00')
average = pd.Timestamp(sum(ts.value for ts in timestamps) / len(timestamps))