<p>您的代码应该更改:</p>
<pre><code>df = pd.read_csv("n2.txt")
g = df.groupby('apn')
f = g.last()
</code></pre>
<p>使用<a href="http://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.Series.to_csv.html" rel="nofollow noreferrer">^{<cd1>}</a>,因为<code>f</code>的输出是<code>Series</code>:</p>
<pre><code>f.to_csv(file)
</code></pre>
<p>或者将<a href="http://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.to_csv.html" rel="nofollow noreferrer">^{<cd4>}</a>与convert <code>index</code>一起使用到2列<code>DataFrame</code>:</p>
<pre><code>f.reset_index().to_csv(file, index=False)
</code></pre>
<p>或与<a href="http://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.drop_duplicates.html" rel="nofollow noreferrer">^{<cd7>}</a>一起使用溶液:</p>
<pre><code>df = pd.read_csv("n2.txt")
df = df.drop_duplicates('apn', keep='last')
df.to_csv(file, index=False)
</code></pre>
<p>在您的解决方案中,使用<code>Index</code>选择<code>index</code>的<code>Series</code>:</p>
<pre><code>for r in f.itertuples(index=True, name='Pandas'):
print(getattr(r,'Index'), getattr(r,'date'))
3704-156 11/22/2019
3732-231 11/15/2019
5515-004 10/23/2019
</code></pre>