如何从另一列中的字符串中获取下划线左右的单词？python

我已将目录中的所有.csv连接到一个大数据框中，其中一列是每个文件的文件名

以下是我的文件名：

['Accelerometer-2011-05-30-09-36-50-brush_teeth-f1.txt',

 'Accelerometer-2011-05-30-08-35-11-brush_teeth-f1.txt',

 'Accelerometer-2011-06-02-10-45-50-brush_teeth-f1.txt',

 'Accelerometer-2011-06-02-10-42-22-brush_teeth-f1.txt',

 'Accelerometer-2011-05-31-15-16-47-brush_teeth-f1.txt',

 'Accelerometer-2011-05-30-21-55-04-brush_teeth-m2.txt',

 'Accelerometer-2011-04-11-13-28-18-brush_teeth-f1.txt',

etc...]

我想创建另一个名为['Action']的列，它将从文件名中提取下划线的左侧和右侧。因此，在这种情况下，它将是“刷牙”

---

我将如何在python中实现这一点？

我建议，根据文件大小，在加载到pandas之前，使用普通python完成所有的争论，因为速度会更快。我的两分钱。这是解决问题的一种方法。我相信评论部分的选项将不仅仅解决这个问题。由于位置的原因，它的索引为-2

text = ['Accelerometer-2011-05-30-09-36-50-brush_teeth-f1.txt',

        'Accelerometer-2011-05-30-08-35-11-brush_teeth-f1.txt',

        'Accelerometer-2011-06-02-10-45-50-brush_teeth-f1.txt',

        'Accelerometer-2011-06-02-10-42-22-brush_teeth-f1.txt',

        'Accelerometer-2011-05-31-15-16-47-brush_teeth-f1.txt',

        'Accelerometer-2011-05-30-21-55-04-brush_teeth-m2.txt',

        'Accelerometer-2011-04-11-13-28-18-brush_teeth-f1.txt',]

(pd.DataFrame(text)
.assign(Action = lambda x: x[0].str.split('-').str[-2]))

       0                                                Action
0   Accelerometer-2011-05-30-09-36-50-brush_teeth-...   brush_teeth
1   Accelerometer-2011-05-30-08-35-11-brush_teeth-...   brush_teeth
2   Accelerometer-2011-06-02-10-45-50-brush_teeth-...   brush_teeth
3   Accelerometer-2011-06-02-10-42-22-brush_teeth-...   brush_teeth
4   Accelerometer-2011-05-31-15-16-47-brush_teeth-...   brush_teeth
5   Accelerometer-2011-05-30-21-55-04-brush_teeth-...   brush_teeth
6   Accelerometer-2011-04-11-13-28-18-brush_teeth-...   brush_teeth

---

我同意这些评论。根据命名模式的稳定性，您根本不需要正则表达式。你可以这样解决：

mylist = ['Accelerometer-2011-05-30-09-36-50-brush_teeth-f1.txt',
 'Accelerometer-2011-05-30-08-35-11-brush_teeth-f1.txt',
 'Accelerometer-2011-06-02-10-45-50-wash_face-f1.txt',
 'Accelerometer-2011-06-02-10-42-22-brush_hair-f1.txt',
 'Accelerometer-2011-05-31-15-16-47-wash_hair-f1.txt',
 'Accelerometer-2011-05-30-21-55-04-iron_clothes-m2.txt',
 'Accelerometer-2011-04-11-13-28-18-make_bed-f1.txt']

output = []
for i in mylist:
    result = i.split("-")
    for z in result:
        if "_" in z:
            output.append(z)

print(output)
>>> ['brush_teeth', 'brush_teeth', 'wash_face', 'brush_hair', 'wash_hair', 'iron_clothes', 'make_bed']

---

我们可以使用正则表达式提取文件扩展名之前的值，该值也不等于使用正向前瞻的破折号

df['file_name'].str.extract(r'(\w+(?=\s*-\w+\.[^\.]))')

---

---

您的所有文件都遵循相同的命名模式吗<代码>“a-b-c-action_name-x-y-z.txt”？@Guimoute是的，它们都遵循相同的命名模式！因此，请使用

split（'-'）[3]

或任何您需要的数字。您可以使用regex

\W（\W+\uw+）\W

如果匹配短语后始终有相同数量的连字符，此解决方案将起作用。如果偶尔出现不同数量的连字符，则会失败。我喜欢lambda函数的优雅。@Matt，谢谢你的反馈。OP说这些文件具有相同的模式，因此解决方案是这样的。

print(df)
                                           file_name       action
0  Accelerometer-2011-05-30-09-36-50-brush_teeth-...  brush_teeth
1  Accelerometer-2011-05-30-08-35-11-brush_teeth-...  brush_teeth
2  Accelerometer-2011-06-02-10-45-50-brush_teeth-...  brush_teeth
3  Accelerometer-2011-06-02-10-42-22-brush_teeth-...  brush_teeth
4  Accelerometer-2011-05-31-15-16-47-brush_teeth-...  brush_teeth
5  Accelerometer-2011-05-30-21-55-04-brush_teeth-...  brush_teeth
6  Accelerometer-2011-04-11-13-28-18-brush_teeth-...  brush_teeth