pandas 基于正则表达式过滤数据框
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/22290000/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
Filtering a dataframe based on a regex
提问by Amelio Vazquez-Reina
Say I have a dataframe my_dfwith a column 'brand', I would like to drop any rows where brand is either toyotaor bmw.
假设我有一个my_df带有列的数据框'brand'，我想删除品牌为toyota或 的任何行bmw。
I thought the following would do it:
我认为以下会做到这一点：
my_regex = re.compile('^(bmw$|toyota$).*$')
my_function = lambda x: my_regex.match(x.lower())
my_df[~df['brand'].apply(my_function)]
but I get the error:
但我收到错误：
ValueError: cannot index with vector containing NA / NaN values
Why? How can I filter my DataFrame using a regex?
为什么？如何使用正则表达式过滤我的 DataFrame？
回答by behzad.nouri
I think re.matchreturns Nonewhen there is no match and that breaks the indexing; below is an alternative solution using pandas vectorized string methods; note that pandas string methods can handle null values:
我认为在没有匹配项时re.match返回None并破坏索引；下面是使用Pandas矢量化字符串方法的替代解决方案；请注意，pandas 字符串方法可以处理空值：
>>> df = pd.DataFrame( {'brand':['BMW', 'FORD', np.nan, None, 'TOYOTA', 'AUDI']})
>>> df
brand
0 BMW
1 FORD
2 NaN
3 None
4 TOYOTA
5 AUDI
[6 rows x 1 columns]
>>> idx = df.brand.str.contains('^bmw$|^toyota$',
flags=re.IGNORECASE, regex=True, na=False)
>>> idx
0 True
1 False
2 False
3 False
4 True
5 False
Name: brand, dtype: bool
>>> df[~idx]
brand
1 FORD
2 NaN
3 None
5 AUDI
[4 rows x 1 columns]
