<p>只需对生成的<code>pd.Series</code>使用<code>numpy.sqrt()</code>(<a href="http://docs.scipy.org/doc/numpy-1.10.1/reference/generated/numpy.sqrt.html" rel="noreferrer">see docs</a>):</p>
<pre><code>import numpy as np
np.sqrt(football[['wins', 'losses']].sum(axis=1))
</code></pre>
<p>但是,当然有几种方法可以达到相同的效果-请参见下面的说明:</p>
<pre><code>df = pd.DataFrame.from_dict(data={'col_1': np.random.randint(low=1, high=10, size=10), 'col_2': np.random.randint(low=1, high=10, size=10)}, orient='index').T
df['sum'] = df[['col_1', 'col_2']].sum(axis=1)
df['np'] = np.sqrt(df[['col_1', 'col_2']].sum(axis=1))
df['apply'] = df[['col_1', 'col_2']].sum(axis=1).apply(np.sqrt)
df['**'] = df[['col_1', 'col_2']].sum(axis=1) ** .5
col_1 col_2 sum np apply **
0 8 3 11 3.316625 3.316625 3.316625
1 4 1 5 2.236068 2.236068 2.236068
2 6 2 8 2.828427 2.828427 2.828427
3 4 1 5 2.236068 2.236068 2.236068
4 4 7 11 3.316625 3.316625 3.316625
5 7 4 11 3.316625 3.316625 3.316625
6 5 5 10 3.162278 3.162278 3.162278
7 1 2 3 1.732051 1.732051 1.732051
8 6 6 12 3.464102 3.464102 3.464102
9 5 7 12 3.464102 3.464102 3.464102
</code></pre>