pandas如何将datetime64[ns]转为字符串日期
作者:呆萌的代Ma
这篇文章主要介绍了pandas如何将datetime64[ns]转为字符串日期,具有很好的参考价值,希望对大家有所帮助。如有错误或未考虑完全的地方,望不吝赐教
将datetime64[ns]转为字符串日期
将datetime64[ns]转为字符串日期(“%Y-%m-%d”)最核心的用法是:
pandas.series.dt.strftime('%Y-%m-%d')
如果是DataFrame或Series的index,则转换最核心的用法是:
pandas.DataFrame.index.strftime("%Y-%m-%d")
示例代码
将series转为字符串日期:
import pandas as pd def convert_datetime(col_series: pd.Series): """series datetime64[ns] 转 字符串日期""" if col_series.dtype == "datetime64[ns]": return col_series.dt.strftime('%Y-%m-%d') else: return col_series def main(): time_series = pd.Series(pd.date_range(start="20200101", periods=20, freq="D")) new_time_series = convert_datetime(time_series) print(time_series, "\n") print(new_time_series) if __name__ == '__main__': main()
使用apply()将整个dataframe的所有datetime64[ns]都转为object类型的日期数据
import pandas as pd import numpy as np def convert_datetime(col_series: pd.Series): """series datetime64[ns] 转 字符串日期""" if col_series.dtype == "datetime64[ns]": return col_series.dt.strftime('%Y-%m-%d') else: return col_series def main(): time_df = pd.DataFrame(index=np.arange(0, 20)) time_df['dt_col'] = pd.date_range(start="20200101", periods=20, freq="D") time_df['num_col'] = np.random.random(size=20) convert_time_df = time_df.apply(convert_datetime, axis=0) print(time_df.dtypes, "\n ==============") print(convert_time_df.dtypes) if __name__ == '__main__': main()
python datetime与字符串、时间戳与字符串相互转换
用flask处理前端传过来的时间参数时,有可能是时间,也有可能是字符串,在不需要前端改动的情况下,后端可以自己处理。
情况1:将datetime形式转为需要的字符串
(这样的字符串在写原生sql语句是可以当作实参传递使用)
import datetime time1 = datetime.datetime.now() print(type(time1)) print(time1)#假设前端传的形式不符合后端要求 time1 = time1.strftime('%Y-%m-%d %H:%M:%S')#只取年月日,时分秒 print(type(time1)) print(time1)
情况2:将字符串形式的时间转为datetime形式
import time,datetime str_time = '2020-9-20 21:33:21' fmt = '%Y-%m-%d %H:%M:%S' print(str_time) print(type(str_time)) str_time = datetime.datetime.strptime(str_time, "%Y-%m-%d %H:%M:%S") print(str_time) print(type(str_time))
情况3:约定前端传过来datetime形式
形如2020-09-20 21:49:58.986521,我们已经将其处理成了字符串,我们只取前端传过来的年月日,后面时分秒由自己添加,这时我们可以使用原生sql语句进行查询,形如(只需要看懂传字符串能查就行)
point_detect = db.session.execute("select a.id as flag_id,a.patrol_time,b.id as point_id,b.point_number,b.x_coor,b.y_coor from pipe_user_point a left join pipe_point b on a.pipe_point_id = b.id and a.pipe_user_id = '%s' and a.patrol_time <= '%s' and a.patrol_time >= '%s' and b.is_active = 1" % (patrol_id,start_time,end_time)).fetchall()
import time,datetime start_time = datetime.datetime.now() print(type(start_time)) print(start_time)#假设前端传的形式不符合后端要求 start_time = start_time.strftime('%Y-%m-%d')#只取年月日 时分秒由自己添加 print(type(start_time)) print(start_time) start_time = start_time + ' 00:00:00' print(type(start_time)) print(start_time) end_time = datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S') print(type(end_time)) print(end_time)
情况4:对datetime形式的时间进行减操作
其中end_time会减去一天
import time,datetime start_time = datetime.datetime.now() print(type(start_time)) print(start_time) end_time = datetime.datetime.now() print(type(end_time)) print(end_time) time = start_time - datetime.timedelta(days=1)#取一天之前 print(type(time)) print(time) time = time.strftime('%Y-%m-%d %H:%M:%S') print(type(time)) print(time)
21-7-7更新
情况5:将前端毫秒时间戳转为年月日时分秒
def time_change_str(int_millisecond_time_stamp): dateArray = datetime.datetime.fromtimestamp(int_millisecond_time_stamp / 1000) otherStyleTime = dateArray.strftime("%Y-%m-%d %H:%M:%S") return otherStyleTime #flask中的model.query.filter(model.report_time >= time1)可行
以上为个人经验,希望能给大家一个参考,也希望大家多多支持脚本之家。