从DataFrame提取指定词生成新列并删除无关行的实现求助

以下是满足需求的可行代码实现：

import pandas as pd

data = {
    "Company": [["ConsenSys"], ["Cognizant"], ["IBM"], ["IBM"], ["Reddit, Inc"], ["Reddit, Inc"], ["IBM"]],
    "skills": [
        ['services', 'scientist technical expertise', 'databases'],
        ['datacomputing tools experience', 'deep learning models', 'cloud services'],
        ['quantitative analytical projects', 'financial services', 'field experience'],
        ['filesystems server architectures', 'systems', 'statistical analysis', 'data analytics', 'workflows', 'aws cloud services'],
        ['aws services'],
        ['data mining statistics', 'statistical analysis', 'aws cloud', 'services', 'data discovery', 'visualization'],
        ['communication skills experience', 'services', 'manufacturing environment', 'sox compliance']
    ]
}

dff = pd.DataFrame(data)

# 定义需要提取的目标关键词
target_words = ['services', 'statistical analysis']

# 生成new_col：提取skills中包含目标词的关键词
dff['new_col'] = dff['skills'].apply(
    lambda skills_list: [word for word in target_words if any(word in skill for skill in skills_list)]
)

# 删除不包含任意目标词的行，重置索引
dff = dff[dff['new_col'].apply(len) > 0].reset_index(drop=True)

print(dff)

代码说明 #

1. 生成new_col：通过apply遍历每一行的skills列表，检查每个目标关键词是否出现在列表的任意元素中（支持子串匹配，比如aws cloud services会匹配到services），将符合条件的关键词收集到new_col中。
2. 过滤无效行：筛选出new_col不为空的行（即至少包含一个目标词），并重置索引。

如果需要精确匹配（仅提取与目标词完全一致的元素），只需将判断条件修改为：

dff['new_col'] = dff['skills'].apply(lambda skills_list: [word for word in target_words if word in skills_list])

内容的提问来源于stack exchange，提问作者hmmmx2