擅长:python、mysql、java
<p>我认为需要转换<a href="http://pandas.pydata.org/pandas-docs/stable/generated/pandas.to_datetime.html" rel="nofollow noreferrer">^{<cd1>}</a>,然后转换为<code>numpy array</code>,并转换为<code>int64</code>,最后除以<code>10**6</code>,因为本机格式是<code>nanosecond</code>s:</p>
<pre><code>df['Time'] = pd.to_datetime(df['Time']).values.astype(np.int64) // 10**6
print (df)
Time User value
0 1522419529600 U101 23
1 1522419529800 U102 22
2 1522419530000 U101 24
</code></pre>