<p>在下面,我首先检查了seconds(<code>df1</code>)不是<code>NaN</code>之后,将其转换为<code>datetime.timedelta</code>对象,然后将这些值添加到<code>df2</code>中的日期。在</p>
<p>在熊猫0.13.1下测试。在</p>
<pre><code>import datetime as dt
import pandas as pd
df1 = pd.DataFrame({'z1': [None, 2.6, None, 0.1, 4.7, 0.1],
'z2': [None, 3.4, None, 1.1, 5.2, 0.6],
'z3': [None, 63, None, 60.7, 64.9, 61.1]})
df2 = pd.DataFrame({'Datetime': ['2014-09-01 05:22', '2014-09-01 05:38', '2014-09-01 06:08',
'2014-09-01 06:27', '2014-09-01 06:37', '2014-09-01 06:57']})
df2['Datetime'] = pd.to_datetime(df2.Datetime)
result = df1.applymap(lambda x: dt.timedelta(seconds=x) if not np.isnan(x) else None)
+ np.tile(df2.values, (1, df1.shape[1]))
>>> pd.DataFrame(result)
z1 z2 z3
0 NaT NaT NaT
1 2014-09-01 05:38:02.600000 2014-09-01 05:38:03.400000 2014-09-01 05:39:03
2 NaT NaT NaT
3 2014-09-01 06:27:00.100000 2014-09-01 06:27:01.100000 2014-09-01 06:28:00.700000
4 2014-09-01 06:37:04.700000 2014-09-01 06:37:05.200000 2014-09-01 06:38:04.900000
5 2014-09-01 06:57:00.100000 2014-09-01 06:57:00.600000 2014-09-01 06:58:01.100000
</code></pre>