datetime使用

 

通过当前日期，获取最近第一个周五、第二个周五、每季度末最后一个周五

    def get_current_week(self, symbol: str, start_date: datetime):
        i = 0
        while start_date.weekday() != 4:
            i = 1
            start_date += timedelta(days=1)
        if i:
            this_week = start_date
        else:
            this_week = start_date + timedelta(days=7)
        next_week = this_week + timedelta(days=7)
        quarter = next_week + timedelta(days=7)
        # 获取日期本季度第一个月(month)和最后一个月(month + 3)
        month = (quarter.month - 1) - (quarter.month - 1) % 3 + 1
        if month + 3 > 12:
            quarter_year = quarter.year + 1
            quarter_month = month + 3 -12
        else:
            quarter_year = quarter.year
            quarter_month = month + 3
        # 获取季度最后一天
        this_quarter_end = datetime(quarter_year, quarter_month, 1) - timedelta(days=1)
        while this_quarter_end - quarter > timedelta(days=7):
            quarter += timedelta(days=7)

        self.week_dic = {'CW': this_week, 'NW': next_week, 'CQ': quarter}
        week = symbol.split('_')[1].split('.')[0]
        return self.week_dic[week]

 

https://docs.python.org/3/library/datetime.html

http://blog.csdn.net/wirelessqa/article/details/7973121/

 

Pandas详解七之DatetimeIndex、PeriodIndex和TimedeltaIndex时间序列

 https://blog.csdn.net/weixin_38168620/article/details/79596564

 

比较好的介绍date、datetime、time、timedelta文章  https://www.cnblogs.com/chengd/articles/7291143.html

 

1、有一个问题，如果有一列，读取出来是如下格式，如何调整为date类型格式，将后面时、分、秒都去掉。

解答：

In [1]: from dateutil.parser import parse

In [2]: parse('2011-01-03')
Out[2]: datetime.datetime(2011, 1, 3, 0, 0)

In [3]: parse('2011-01-03').date()
Out[3]: datetime.date(2011, 1, 3)

 parse('2011-01-20 23:10:33')
Out[6]: datetime.datetime(2011, 1, 20, 23, 10, 33)

In [7]: parse('2011-01-20 23:10')
Out[7]: datetime.datetime(2011, 1, 20, 23, 10)

In [8]: parse('2011-01-20 23:10').date()
Out[8]: datetime.date(2011, 1, 20)

3、通过tushare获得交易日时间序列，问题一：获取到每个月第三个星期一，如果没有顺延到下一个交易日。问题二：获取每个月最后一个交易日。

In [1]: import tushare as ts
In [3]: ts.trade_cal()
Out[3]:
      calendarDate  isOpen
0       1990-12-19       1
1       1990-12-20       1
2       1990-12-21       1
3       1990-12-22       0