MATLAB datenum日期转换为Python日期

摘要

MATLAB datenum时间格式参数众多，本文只简单关注 units 参数，即基准年份和计时度量（天、小时）。

命令行演示在 ipython 和 Octave 中进行。

示例1：小时制，基准年份1800-1-1

Time Attributes:
                     units              = 'hours since 1800-1-1 00:00:0.0'
                     long_name          = 'Time'
                     axis               = 'T'
                     standard_name      = 'time'
                     coordinate_defines = 'start'
                     delta_t            = '0000-00-01 00:00:00'
                     actual_range       = [1.74e+06 1.75e+06]
                     avg_period         = '0000-00-01 00:00:00'
time(1:5):
     1744392
     1744416
     1744440
     1744464
     1744488
#来源： 参考1

# python
import numpy as np
origin = np.datetime64('1800-01-01', 'D')
date1 = 1744392/24 * np.timedelta64(1, 'D') + origin   # '1999-01-01'

示例2：一天制，基准年份1800-1-1

time {
        String units "days since 1800-1-1 00:00:00";
        String long_name "Time";
        Float64 actual_range 19723.00000000000, 76214.00000000000;
        String delta_t "0000-01-00 00:00:00";
        String avg_period "0000-01-00 00:00:00";
        String prev_avg_period "0000-00-07 00:00:00";
        String standard_name "time";
        String axis "t";
}
#来源： 参考2

# python
import numpy as np
origin = np.datetime64('1800-01-01', 'D')
date2 = 19723 * np.timedelta64(1, 'D') + origin   # '1854-01-01'

示例3：时间列表转换

import numpy as np
import pandas as pd
datenums = np.array([730990, 733301, 729159,  734471, 736858, 731204]) # 来源：参考7
timestamps = pd.to_datetime(datenums-719529, unit='D')

'''
DatetimeIndex(['2001-05-19', '2007-09-16', '1996-05-14',
               '2010-11-29', '2017-06-12', '2001-12-19'],
              dtype='datetime64[ns]', freq=None)
'''