pandas 拆分列的一个特别例子

柔光的暖阳◎ 2022-12-22 13:25 139阅读 0赞

class\_name列同时包含课程名称和同类群组编号。  
我想将栏分为两栏（名称，同类群组编号）  
  
从：

（我知道它应该像1st，2nd，3rd一样，但是字符串使用的是我的语言，并且在数字之后我们反复使用相同的字符）。  
  
至：

| class_name | class_cohort | | introduction to programming | 1 | | introduction to programming | 2 | | introduction to programming | 3 | | introduction to programming | 4 | | algorithms and data structure | 1 | | alrogithms and data structure | 2 | | object-oriented programming | 1 | | database systems | 1 |

这是我一直在努力的代码：

import pandas as pd course_count = 100 df = pd.read_csv("course.csv", nrows=course_count) cols_interest=['class_name', 'class_department', 'class_type', 'student_target', 'student_enrolled'] df = df[cols_interest] df.insert(1, 'class_cohort', 0) # this is how I extract the numbers df['class_name'].str.extract('(\d)').head() # but I cannot figure out a way to copy those values into column 'class_cohort' which I filled with 0's. # once I figure that out, I plan to discard the last digits df['class_name'] = df['class_name'].map(lambda x: str(x)[:-1])

我简要地研究了一种解决方案，将逗号放在第1，第2，第3之前，然后使用逗号作为分隔符来拆分列，但我想不出一种方法来替换所有数字\\ s1th->，1th。

**最佳答案**

您可以[indexing by positions][]：

df['class_cohort'] = df['class_name'].str[-3:-2] df['class_name'] = df['class_name'].str[:-4] print df class_name class_cohort 0 cs101 1 1 cs101 2 2 cs101 3 3 cs101 4 4 algorithms 1 5 algorithms 2

[我的博客][Link 1]  
或使用[`str.extract`][str.extract]：

df['class_cohort'] = df['class_name'].str.extract('(\d)') df['class_name'] = df['class_name'].str[:-4] print df class_name class_cohort 0 introduction to programming 1 1 introduction to programming 2 2 introduction to programming 3 3 introduction to programming 4 4 algorithms and data structure 1 5 algorithms and data structure 2

[indexing by positions]: http://pandas.pydata.org/pandas-docs/stable/text.html#indexing-with-str
[Link 1]: https://blog.csdn.net/dudu3332
[str.extract]: http://pandas.pydata.org/pandas-docs/stable/generated/pandas.Series.str.extract.html