字符串清理函数产生意外输出

Question

我希望清理以下数据帧的配置文件列：

name    profile
6   Pedro   ["\n Design ...
7   Leonardo    ["\n Design ...
8   Daniel  ["\n JavaScript ...
9   Mario   ["\n JavaScript ...
10  Christi     ["\n Design ...

我已经在单独的行上测试了以下函数...

def clean_profile(row):
    for index, row in new_df2["profile"].items():
        str_row = str(row)
        clean_row = (
            '""'.join(str_row)
            .replace(",", "")
            .replace('""', "")
            .replace("\\n                    ", "")
            .replace("                ", "")
        )
    return clean_row

...and找到它来转换这个字符串：

'["\\n                    Design                ","\\n                    Design                "]'

添加到这个清理过的字符串：

'["Design","Design"]'

(额外的替换方法对于清理非常杂乱的字符串是必要的，就像下面这样：)

'{"Tools ""    Google Analytics            ":null,"    Google Adsense            ":null,"    MailChimp            ":null,"    Google Adwords            ","Containers ""    Docker            ","Digital ""    SEO            ":null,"    Email Marketing            ":null,"    Article Writing            ":null,"    Market Research            ":null,"    Social Media            ":null,"    Inbound Marketing            ","*Nix ""    Ubuntu            ":null,"    Linux            ","Java ""    Java    ","Python ""    Django            ":null,"    Python    ":null,"    Flask            ","Databases ""    MySQL Management            ":null,"    MongoDB Management            ":null,"    PostgreSQL Management            ","Visual ""    Brand Design            ":null,"    Graphic Design            ":null,"    Logo Design            ","HTML ""    HTML    ","Version Control ""    Git            ","PHP ""    Laravel            ":null,"    Wordpress            ":null,"    PHP    ":null,"    Symfony            ","Mobile ""    React Native            ","Ruby ""    Ruby    ":null,"    Sinatra            ":null,"    Rails            ","Project Management ""    Agile Methodology            ":null,"    Client Management            ":null,"    Scrum            ","English ""    Written English    ":null,"    Spoken English            ","Configuration Management ""    Chef            ","Webserver ""    Nginx            ":null,"    Apache            ","CDN ""    AWS CloudFront            ":null,"    Cloudflare            ","Other ""    C++            ","Experience ""    Creative Direction            ":null,"    UI/UX Design            ":null,"    Wireframing            ","JavaScript ""    JavaScript    ":null,"    TypeScript            ":null,"    Redux            ":null,"    Angular JS            ":null,"    Angular            ":null,"    D3.js            ":null,"    Node.js            ":null,"    React            ":null,"    Flux            ":null,"    Express            ","CSS ""    SASS            ":null,"    LESS            ":null,"    CSS    ","Hosting ""    Heroku            ":null,"    Digital Ocean            ":null,"    AWS            ","Automated Testing ""    TDD            ":null,"    Automated Testing    ":null,"    BDD            ":null,"    Jest            ","Traditional ""    Outbound Marketing            ":null,"    Brand Strategy            ","Data Science ""    Data Science    ":null,"    Data Analysis            ":null,"    Machine Learning            ":null,"    Data Visualization            ":null,"    R            ":null,"    Statistics            "}'

当我循环遍历dataframe的所有行时，我得到对所有行重复的结果：

["JavaScriptDevOpsPHPJavaScriptDevOpsPHP"]

或者这样：

<function clean_profile at 0x0845CB20>

我已经尝试了一些不同的方法，但是worked...is没有人能够解释这里发生了什么，也许可以建议一个更好的方法来清理这些字符串？

谢谢!

Answer 1

看起来您没有在for循环外部声明初始clean_row变量，因此您的clean_row将始终等于您上次清除的任何字符串。

def clean_profile(row):
    clean_row = "" //added this line
    for index, row in new_df2["profile"].items():
        str_row = str(row)
        clean_row = (
            '""'.join(str_row)
            .replace(",", "")
            .replace('""', "")
            .replace("\\n                    ", "")
            .replace("                ", "")
        )
    return clean_row

此外，我还会研究一下用于字符串清理的strip函数。这里有一个很好的例子：https://www.geeksforgeeks.org/clean-the-string-data-in-the-given-pandas-dataframe/