Setting working shifts ammong days in a Pandas dataframe

user_6634159 0
Viewed 128

I have a Pandas dataframe with datetime64 column.

               date
2022-01-30 10:57:00
2022-01-30 12:02:00
2022-01-30 18:58:00
2022-01-30 19:27:00
2022-01-30 12:00:00
2022-01-31 02:10:00
2022-01-31 01:55:00
2022-01-31 01:24:00
2022-01-31 02:37:00
...

What I need is to set working shifts (I don't know if shift is correct for turno in spanish...) with this conditions

  • Night shift: 22:00:00 - 05:00:00
  • Morning shift: 05:00:00 - 13:00:00
  • Afternoon shift: 13:00:00 - 22:00:00

to store it in a new column:

               date shift
2022-01-30 10:57:00 Morning
2022-01-30 12:02:00 Morning
2022-01-30 18:58:00 Afternoon 
2022-01-30 19:27:00 Afternoon 
2022-01-30 12:00:00 Morning
2022-01-31 02:10:00 Night 
2022-01-31 01:55:00 Night 
2022-01-31 01:24:00 Night
...

I have done this with this code (it works fine).

startTime = pd.to_datetime(entradaDf['startDate']).dt.strftime('%H:%M:%S')

condShift = [
    startTime.ge('05:00:00') & startTime.lt('13:30:00'),  # Morning
    startTime.ge('13:30:00') & startTime.lt('22:00:00'),  # Afternoon
    startTime.ge('22:00:00') & startTime.lt('00:00:00'),  # Night
    startTime.ge('00:00:00') & startTime.lt('05:00:00'),  # Night
]

shiftValues = ['Mañana', 'Tarde', 'Noche', 'Noche']

df['shift'] = np.select(condShift , shiftValues , default='Night')

It work fine but, is there a way to do this more efficiently?

pandas dataframe datetime python
3 Answers

You can use np.select:

import numpy as np
from datetime import time

df['date'] = pd.to_datetime(df['date'])

condlist = [df['date'].dt.time.between(time(5), time(13, 30), inclusive='left'),
            df['date'].dt.time.between(time(13, 30), time(22), inclusive='left')]

choicelist = ['Mañana', 'Tarde']

df['shift'] = np.select(condlist=condlist, choicelist=choicelist, default='Noche')

Output:

>>> df
                 date   shift
0 2022-01-30 10:57:00  Mañana
1 2022-01-30 12:02:00  Mañana
2 2022-01-30 18:58:00   Tarde
3 2022-01-30 19:27:00   Tarde
4 2022-01-30 12:00:00  Mañana
5 2022-01-31 02:10:00   Noche
6 2022-01-31 01:55:00   Noche
7 2022-01-31 01:24:00   Noche
8 2022-01-31 02:37:00   Noche
user_152399510

You can use cut on the time part of the dates, converted to seconds from midnight:

df['shift'] = pd.cut(df.date.dt.hour * 3600 + df.date.dt.minute * 60 + df.date.dt.second,
                     [0, 5 * 3600, (13 * 60 + 30) * 60 , 22 * 3600, 24 * 3600],
                     right=False,
                     labels=['Noche', 'Mañana', 'Tarde', 'Noche'],
                     ordered=False)

Result:

                 date   shift
0 2022-01-30 10:57:00  Mañana
1 2022-01-30 12:02:00  Mañana
2 2022-01-30 18:58:00   Tarde
3 2022-01-30 19:27:00   Tarde
4 2022-01-30 12:00:00  Mañana
5 2022-01-31 02:10:00   Noche
6 2022-01-31 01:55:00   Noche
7 2022-01-31 01:24:00   Noche
8 2022-01-31 02:37:00   Noche
user_39443220

I always use pandas' apply() function to transform a row or column into some new row or column. You can use something like this:

from io import StringIO
import pandas as pd
import numpy as np

# Create your data frame
input_string = """startDate
2022-01-30 10:57:00
2022-01-30 12:02:00
2022-01-30 18:58:00
2022-01-30 19:27:00
2022-01-30 12:00:00
2022-01-31 02:10:00
2022-01-31 01:55:00
2022-01-31 01:24:00
2022-01-31 02:37:00"""

data = StringIO(input_string)
entradaDf = pd.read_csv(data, sep='  ', engine='python')

# Create series with time only
time_series = pd.to_datetime(entradaDf['startDate']).dt.strftime('%H:%M:%S')

# Function defining what shift we have
def defineShift(startTime):
    if startTime >= '05:00:00' and startTime < '13:30:00':
        return 'Mañana'
    if startTime >= '13:30:00' and startTime < '22:00:00':
        return 'Tarde'
    return 'Noche'

# Apply transformation of time to shift and save in the df
entradaDf['newShift'] = time_series.apply(lambda x: defineShift(x))
entradaDf

Output:

startDate   newShift
0   2022-01-30 10:57:00 Mañana
1   2022-01-30 12:02:00 Mañana
2   2022-01-30 18:58:00 Tarde
3   2022-01-30 19:27:00 Tarde
4   2022-01-30 12:00:00 Mañana
5   2022-01-31 02:10:00 Noche
6   2022-01-31 01:55:00 Noche
7   2022-01-31 01:24:00 Noche
8   2022-01-31 02:37:00 Noche
user_79479940
Related