how to convert 1 digit, 2 digit, 3 digit, 4 digit number to '%H%M'?

Question

I'm using Fedex Dataset from kaggle . There is a column by the name of Actual_Shipment_Time which contains numbers from single digit to four digits which I'm trying to convert to time.

For example

• 5 needs to be 00:05,
• 23 needs to be 00:23
• 345 needs to be 03:45
• 2145 needs to be 21:45

Even 12 hours format is also acceptable.

When I run this code

df['Actual_Shipment_Time'] = pd.to_datetime(df['Actual_Shipment_Time'], format = '%H%M').dt.strftime('%H%M')

I get this error:

ValueError: time data '9' does not match format '%H%M' (match)

Answer 1

You can try zfill the column to 4 numbers

df['Actual_Shipment_Time'] = (pd.to_datetime(df['Actual_Shipment_Time']
                                             .astype(str)
                                             .str.zfill(4), format='%H%M')
                              .dt.strftime('%H:%M'))

print(df)

  Actual_Shipment_Time
0                00:05
1                00:23
2                03:45
3                21:45

There is no such time as 24:00 , so Pandas gives following error

$ pd.to_datetime('2400', format='%H%M')

ValueError: unconverted data remains: 0

To overcome it, we can use errors='coerce'

df['Actual_Shipment_Time'] = (pd.to_datetime(df['Actual_Shipment_Time']
                                             .fillna(0).astype(int).astype(str)  # fill nan value by 0 and convert column to int then to string
                                             .str.zfill(4), format='%H%M', errors='coerce')
                              .dt.strftime('%H:%M')
                              .fillna('00:00')) # fill nat value by 00:00, nat is from 24:00