将函数应用于pandas df中的行 [英] Apply function over rows in pandas df

问题描述

在熊猫df中，我有一列 ['name'] ，其中包含各种操作系统分类，例如'Windows 7'，'Windows 10'，'Linux'，'MobileiOS 9.1"，"OS X 10.12" 等.这是字符串.

In a pandas df, I have a column ['name'] with various Operating System classifications such as 'Windows 7', 'Windows 10', 'Linux', 'Mobile iOS 9.1', 'OS X 10.12', etc. That are strings.

我希望使用此功能来创建新列 ['type'] ，这将是更通用的版本:

I am hoping to use this function to create a new column ['type'] that will be a more generalized version:

def name_group(row):
    if 'Windows' in row:
        name = 'Microsoft Windows'
    elif 'iOS' in row:
        name = 'Apple iOS'
    elif 'OS X' in row:
        name ='Apple Macintosh'
    elif 'Macintosh' in row:
        name = 'Apple Macintosh'
    elif 'Linux' in row:
        name = 'GNU/Linux'
    else:
        name = 'Other'
    return name

当我通过传入单个字符串变量来测试函数时，它可以正常工作，但是由于某种原因，当我将函数应用于df时，它仅对每一行返回"other".

It works correctly when I test the function by passing in a single string variable, but for some reason when I apply the function to the df like this, it only returns "other" for each row.

new_df['type'] = new_df.apply(name_group, axis=1)

对造成这种情况的任何想法?

Any thoughts on what could be causing this?

推荐答案

您需要使用

You need pass column name with Series.apply:

new_df['type'] = new_df['name'].apply(name_group)

但是，如果要使用 DataFrame.应用 ，然后还需要 lambda 函数并传递列名:

But if want use DataFrame.apply then need lambda function and pass name of column too:

new_df['type'] = new_df.apply(lambda x: name_group(x['name']), axis=1)

这篇关于将函数应用于pandas df中的行的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！