如何将netCDF转换后的datetime数组拆分日期与秒数

Hey there! Let's break down how to split your numpy array of datetime objects into separate date components and seconds since midnight. Here are two straightforward approaches depending on your needs:

Approach 1: Using Python Datetime Methods (Simple & Readable) #

This method leverages built-in datetime properties and works well if you prefer explicit, easy-to-follow code:

import numpy as np
from datetime import datetime

# Your converted datetime array (from netCDF)
time_array = np.array([
    datetime(2012, 1, 1, 0, 0), 
    datetime(2012, 1, 1, 6, 0), 
    datetime(2012, 1, 1, 12, 0),
    datetime(2012, 12, 31, 18, 0)
])

# Extract just the date part (as datetime.date objects)
dates = np.array([dt.date() for dt in time_array])

# Calculate seconds since midnight for each entry
seconds_since_midnight = np.array([
    dt.hour * 3600 + dt.minute * 60 + dt.second 
    for dt in time_array
])

Output Notes: #

• dates will be an array of datetime.date objects (e.g., datetime.date(2012, 1, 1)).
• seconds_since_midnight will be integers like 0, 21600 (6 hours), 43200 (12 hours), etc.

Approach 2: Vectorized Numpy Operations (Efficient for Large Arrays) #

If you're working with bigger datasets, this vectorized method avoids loops and runs faster:

import numpy as np

# Convert your datetime array to numpy's native datetime64 format
time_np = np.array(time_array, dtype='datetime64[ns]')

# Extract dates (as datetime64[D] objects, which represent whole days)
dates_np = time_np.astype('datetime64[D]')
# Optional: Convert to Python datetime.date objects if needed
dates = dates_np.astype(object)

# Calculate seconds since midnight using timedelta arithmetic
time_deltas = time_np - dates_np
seconds_since_midnight = time_deltas.astype('timedelta64[s]').astype(int)

Bonus: Format Dates as Strings #

If you need dates in a human-readable string format (like 2012-01-01), add this line:

date_strings = np.array([dt.strftime('%Y-%m-%d') for dt in dates])

内容的提问来源于stack exchange，提问作者Allen Zhang