擅长:python、mysql、java
<p>我没有安装pandas,但是解决方案应该适用于<code>df</code>对象。你知道吗</p>
<pre><code>string="48925661,Y2,9-10 months,7+"
"_".join(re.findall(r'\b\d+\b', string.split(",")[2]))
</code></pre>
<p>试验结果:</p>
<pre><code>>>> "_".join(re.findall(r'\b\d+\b', string.split(",")[2]))
'9_10'
</code></pre>
<p>python脚本:</p>
<pre><code>$ cat test.py
with open("sample.csv") as inputs:
next(inputs) # skip the first line
for line in inputs:
parts = line.strip().split(",")
parts[2] = "_".join(re.findall(r'\b\d+\b', parts[2]))
print(",".join(parts))
</code></pre>
<p>结果:</p>
<pre><code>$python test.py
48925661,Y2,9_10,7+
90764620,Y3,8_9,3
61221204,Y1,2_3,1
</code></pre>