拆分列并格式化列值答案

【问题标题】：Split column and format the column values拆分列并格式化列值
【发布时间】：2017-03-21 04:36:51
【问题描述】：

我正在尝试格式化一列数据。我可以找到拆分列的选项，因为它之间有,，但我无法按输出所示格式化它。

输入

    TITLE,Issn
NATURE REVIEWS MOLECULAR CELL BIOLOGY,"ISSN 14710072, 14710080"
ANNUAL REVIEW OF IMMUNOLOGY,"ISSN 07320582, 15453278"
NATURE REVIEWS GENETICS,"ISSN 14710056, 14710064"
CA - A CANCER JOURNAL FOR CLINICIANS,"ISSN 15424863, 00079235"
CELL,"ISSN 00928674, 10974172"
ANNUAL REVIEW OF ASTRONOMY AND ASTROPHYSICS,"ISSN 15454282, 00664146"
NATURE REVIEWS IMMUNOLOGY,"ISSN 14741741, 14741733"
NATURE REVIEWS CANCER,ISSN 1474175X
ANNUAL REVIEW OF BIOCHEMISTRY,"ISSN 15454509, 00664154"
REVIEWS OF MODERN PHYSICS,"ISSN 00346861, 15390756"
NATURE GENETICS,ISSN 10614036

将 issn 列拆分为两列，因为它有 ,
仅从列中删除单词 ISSN
留下数字 4 位数字后放一个-

预期输出是

    TITLE,Issn
NATURE REVIEWS MOLECULAR CELL BIOLOGY,1471-0072, 1471-0080
ANNUAL REVIEW OF IMMUNOLOGY,0732-0582, 1545-3278
NATURE REVIEWS GENETICS,1471-0056, 1471-0064
CA - A CANCER JOURNAL FOR CLINICIANS,1542-4863, 0007-9235
CELL,0092-8674, 1097-4172
ANNUAL REVIEW OF ASTRONOMY AND ASTROPHYSICS,1545-4282, 0066-4146
NATURE REVIEWS IMMUNOLOGY,1474-1741, 1474-1733
NATURE REVIEWS CANCER, 1474-175X
ANNUAL REVIEW OF BIOCHEMISTRY,1545-4509, 0066-4154
REVIEWS OF MODERN PHYSICS,0034-6861, 1539-0756
NATURE GENETICS,1061-4036

对熊猫的任何建议表示赞赏..提前感谢

更新：
尝试运行答案中提到的两个程序时

import pandas as pd
import re

df = pd.read_csv('new_journal_list.csv', header='TITLE,Issn')

'''
df_split_num = df['Issn'].map(lambda x: x.split('ISSN ')[1].split(', '))
df_dash_num = df_split_num.map(lambda x: [num[:4] + '-' + num[4:] for num in x])

df_split_issn = pd.DataFrame(data=list(df_dash_num), columns=['Issn1', 'Issn2'])
df[['Issn1', 'Issn2']] = df_split_issn
del df['Issn']

print df

'''

df[['Issn1','Issn2']] = (df.pop('Issn').str.extract('ISSN\s+([^,]+),?\s?(.*)', expand=True)
                   .apply(lambda x: x.str[:4]+'-'+x.str[4:]).replace(r'^-$', '', regex=True))

print df

在默认 python 2.7 中运行时，无论哪种情况，我都会收到以下错误

Traceback (most recent call last):
  File "clean_journal_list.py", line 1, in <module>
    import pandas as pd
  File "/usr/local/lib/python2.7/dist-packages/pandas/__init__.py", line 25, in <module>
    from pandas import hashtable, tslib, lib
  File "pandas/src/numpy.pxd", line 157, in init pandas.hashtable (pandas/hashtable.c:38364)

在 python 3.4 中运行时，会看到以下给出的错误

File "clean_journal_list.py", line 21
    print df
           ^
SyntaxError: invalid syntax

【问题讨论】：

您能否将您的数据发布为print(df.to_csv()) 的输出 - 因为它很难解析...
是的，我可以.. ..
@MaxU 完成将输入更改为 csv ..
更改：print df --> print(df) 用于 Python 3.x
是的，我做到了..再次感谢..

标签： python csv pandas dataframe data-cleaning

【解决方案1】：

您需要为此添加一些错误处理，并将其包装在逐行迭代中，但这里是它的要点：

leader, issns = line.split(" ISSN ")
numbers = issns.split(", ")

print leader, ', '.join([ num[:4] + '-' + num[4:] for num in numbers])

关键是将每一行拆分为“ISSN 编号”和“其他所有内容”，然后将 ISSN 编号彼此分开并重新格式化。

【讨论】：

【解决方案2】：

首先，拆分数字并添加破折号。使用方便的地图功能：

df_split_num = df['Issn'].map(lambda x: x.split('ISSN ')[1].split(', '))
df_dash_num = df_split_num.map(lambda x: [num[:4] + '-' + num[4:] for num in x])

接下来，使用拆分出来的 issn 编号创建一个新数据框，并将其放回原始数据框：

df_split_issn = pd.DataFrame(data=list(df_dash_num), columns=['Issn1', 'Issn2'])
df[['Issn1', 'Issn2']] = df_split_issn
del df['Issn']

【讨论】：

能否请您检查更新的输入和输出。排名不再是一列
我假设您的输入是一个数据框。如果 Rank 不是索引，则只需删除上面的 reset_index 和 set_index 行。
非常感谢......但我毫不怀疑它会影响列中值的顺序吗？因为这些是第一列中期刊标题的 ISSN 号？
最后一件事..我将打印相同的df，我没看错？

【解决方案3】：

IIUC 你可以使用Series.str.extract()、apply() 和replace() 方法来做到这一点：

In [33]: df
Out[33]:
                                          TITLE                     Issn
0         NATURE REVIEWS MOLECULAR CELL BIOLOGY  ISSN 14710072, 14710080
1                   ANNUAL REVIEW OF IMMUNOLOGY  ISSN 07320582, 15453278
2                       NATURE REVIEWS GENETICS  ISSN 14710056, 14710064
3          CA - A CANCER JOURNAL FOR CLINICIANS  ISSN 15424863, 00079235
4                                          CELL  ISSN 00928674, 10974172
5   ANNUAL REVIEW OF ASTRONOMY AND ASTROPHYSICS  ISSN 15454282, 00664146
6                     NATURE REVIEWS IMMUNOLOGY  ISSN 14741741, 14741733
7                         NATURE REVIEWS CANCER            ISSN 1474175X
8                 ANNUAL REVIEW OF BIOCHEMISTRY  ISSN 15454509, 00664154
9                     REVIEWS OF MODERN PHYSICS  ISSN 00346861, 15390756
10                              NATURE GENETICS            ISSN 10614036

In [34]: df[['Issn1','Issn2']] = (df.pop('Issn')
    ...:                            .str.extract('ISSN\s+([^,]+),?\s?(.*)', expand=True)
    ...:                            .apply(lambda x: x.str[:4]+'-'+x.str[4:])
    ...:                            .replace(r'^-$', '', regex=True))
    ...:

In [35]: df
Out[35]:
                                          TITLE      Issn1      Issn2
0         NATURE REVIEWS MOLECULAR CELL BIOLOGY  1471-0072  1471-0080
1                   ANNUAL REVIEW OF IMMUNOLOGY  0732-0582  1545-3278
2                       NATURE REVIEWS GENETICS  1471-0056  1471-0064
3          CA - A CANCER JOURNAL FOR CLINICIANS  1542-4863  0007-9235
4                                          CELL  0092-8674  1097-4172
5   ANNUAL REVIEW OF ASTRONOMY AND ASTROPHYSICS  1545-4282  0066-4146
6                     NATURE REVIEWS IMMUNOLOGY  1474-1741  1474-1733
7                         NATURE REVIEWS CANCER  1474-175X
8                 ANNUAL REVIEW OF BIOCHEMISTRY  1545-4509  0066-4154
9                     REVIEWS OF MODERN PHYSICS  0034-6861  1539-0756
10                              NATURE GENETICS  1061-4036

【讨论】：