我有一个包含大约50个特征的数据框。我的实验中涉及到一个分类问题，因此我想通过“GradientBoostingClassifier”来训练模型。该数据框（mydata）被视为训练集。其中一个特征（feature20）是日期，我需要在训练集中也考虑这个特征，因此我尝试将日期转换为datetime64，如下所示：
mydata['feature20']=pd.to_datetime(mydata['feature20'])现在，当我尝试用分类器训练模型时，出现了以下错误：
float() argument must be a string or a number, not 'Timestamp'有什么解决这个问题的想法吗？
回答：
你可以轻松地将日期转换为整数：df[“feature20”].astype(“int64”) // 10**9.
注意： 除非你正在处理时间序列，否则最好不要直接使用datetime特征。通常，你会希望从datetime中提取更多的信息——比如星期几、月份的哪一天、年度的第几周、月份编号等
演示：
In [9]: df = pd.DataFrame({'feature20':pd.date_range('2010-01-01', periods=10)})In [10]: df["new"] = df["feature20"].astype("int64") // 10**9In [11]: dfOut[11]: feature20 new0 2010-01-01 12623040001 2010-01-02 12623904002 2010-01-03 12624768003 2010-01-04 12625632004 2010-01-05 12626496005 2010-01-06 12627360006 2010-01-07 12628224007 2010-01-08 12629088008 2010-01-09 12629952009 2010-01-10 1263081600In [12]: df["date"] = pd.to_datetime(df["new"], unit="s")In [13]: dfOut[13]: feature20 new date0 2010-01-01 1262304000 2010-01-011 2010-01-02 1262390400 2010-01-022 2010-01-03 1262476800 2010-01-033 2010-01-04 1262563200 2010-01-044 2010-01-05 1262649600 2010-01-055 2010-01-06 1262736000 2010-01-066 2010-01-07 1262822400 2010-01-077 2010-01-08 1262908800 2010-01-088 2010-01-09 1262995200 2010-01-099 2010-01-10 1263081600 2010-01-10如果你的日期精确到微秒：
In [28]: df = pd.DataFrame({'feature20':pd.date_range('2010-01-01 01:01:01.123456', freq="123S", periods=10)})In [29]: dfOut[29]: feature200 2010-01-01 01:01:01.1234561 2010-01-01 01:03:04.1234562 2010-01-01 01:05:07.1234563 2010-01-01 01:07:10.1234564 2010-01-01 01:09:13.1234565 2010-01-01 01:11:16.1234566 2010-01-01 01:13:19.1234567 2010-01-01 01:15:22.1234568 2010-01-01 01:17:25.1234569 2010-01-01 01:19:28.123456In [30]: df["new"] = df["feature20"].astype("int64") // 10**3In [31]: dfOut[31]: feature20 new0 2010-01-01 01:01:01.123456 12623076611234561 2010-01-01 01:03:04.123456 12623077841234562 2010-01-01 01:05:07.123456 12623079071234563 2010-01-01 01:07:10.123456 12623080301234564 2010-01-01 01:09:13.123456 12623081531234565 2010-01-01 01:11:16.123456 12623082761234566 2010-01-01 01:13:19.123456 12623083991234567 2010-01-01 01:15:22.123456 12623085221234568 2010-01-01 01:17:25.123456 12623086451234569 2010-01-01 01:19:28.123456 1262308768123456In [32]: df["date"] = pd.to_datetime(df["new"], unit="us")In [33]: dfOut[33]: feature20 new date0 2010-01-01 01:01:01.123456 1262307661123456 2010-01-01 01:01:01.1234561 2010-01-01 01:03:04.123456 1262307784123456 2010-01-01 01:03:04.1234562 2010-01-01 01:05:07.123456 1262307907123456 2010-01-01 01:05:07.1234563 2010-01-01 01:07:10.123456 1262308030123456 2010-01-01 01:07:10.1234564 2010-01-01 01:09:13.123456 1262308153123456 2010-01-01 01:09:13.1234565 2010-01-01 01:11:16.123456 1262308276123456 2010-01-01 01:11:16.1234566 2010-01-01 01:13:19.123456 1262308399123456 2010-01-01 01:13:19.1234567 2010-01-01 01:15:22.123456 1262308522123456 2010-01-01 01:15:22.1234568 2010-01-01 01:17:25.123456 1262308645123456 2010-01-01 01:17:25.1234569 2010-01-01 01:19:28.123456 1262308768123456 2010-01-01 01:19:28.123456