计算 pandas 行中值的外观 [英] Count appearance of a value in a pandas row

问题描述

嘿，我目前正在熊猫中的一列中进行迭代. 现在，我想处理一个值在该列中只出现一次的情况，而不是在它出现多次时的情况. 我尝试了几种方法，但是都没有用. 现在我得到了错误: ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all(). 我只是尝试首先将所有带有id i的行过滤到df_path_counter中，然后我将其视为if子句中的行.我不知道为什么它不起作用. 有任何想法吗? 这是我的代码:

Hey I am currently iterating through a column in pandas. Now i want to handle the case that a value appears only once in that column different than when it appears multiple times. I tried several approaches but none worked. Right now I get the error: ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all(). I just try to filter first all rows with the id i into df_path_counterand than i count it is rows in the if clause. I don't know why it does not work. Any ideas? This is my code:

df_path = pd.DataFrame([(1, 'Germany'),
               (1, 'France'),
               (1, 'Indonesia'),
               (1, 'France'),
               (2, 'France'),
               (1, 'Germany'),
               (1, 'UK'),
               ],
              columns=['id', 'country']
for i, g in df_path.groupby('id'):
    df_path_counter=df_path.loc[df_path['id'] == i]
    if(df_path_counter.count()<=1):
         #...do sth

推荐答案

根据建议，使用Series.value_counts创建一个count == 1的国家/地区列表，并使用带有Series.isin的布尔索引进行过滤:

As suggested, use Series.value_counts to create a list of country with count == 1 and use boolean indexing with Series.isin to filter:

country_counts = df_path['country'].value_counts()
country_1 = country_counts[country_counts.eq(1)].index

df_path[df_path['country'].isin(country_1)]

[出]

    id  country
2   1   Indonesia
6   1   UK

这篇关于计算 pandas 行中值的外观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！