擅长:python、mysql、java
<p>在我看来,使用<a href="http://pandas.pydata.org/pandas-docs/stable/generated/pandas.to_datetime.html" rel="nofollow">^{<cd2>}</a>将字符串转换为<code>datetime</code>,这样就可以对其执行算术运算,如果需要年份或任何其他日期/时间组件,则可以使用矢量化的<a href="http://pandas.pydata.org/pandas-docs/stable/api.html#datetimelike-properties" rel="nofollow">^{<cd3>}</a>访问器:</p>
<pre><code>In [23]:
df['date'] = pd.to_datetime(df['date'])
df
Out[23]:
date
0 2015-01-30
1 2015-01-30
2 2015-01-30
3 2015-01-30
4 2015-01-30
5 2015-01-30
6 2015-01-30
7 2015-01-30
8 2015-01-30
9 2015-01-30
In [24]:
df['year'] = df['date'].dt.year
df
Out[24]:
date year
0 2015-01-30 2015
1 2015-01-30 2015
2 2015-01-30 2015
3 2015-01-30 2015
4 2015-01-30 2015
5 2015-01-30 2015
6 2015-01-30 2015
7 2015-01-30 2015
8 2015-01-30 2015
9 2015-01-30 2015
</code></pre>