如何过滤 pandas 数据框中每一行的元组列表？

Question

您好，我正在尝试过滤第二个元素以 'V' 开头的元组列表，以清理我的数据框。

我有一个 pandas 数据帧调用 'df_my_string' 就像：

样本是：

verbs_tokens
[('[', 'NNS'), ("'Europe", "''"), ('was', 'VBD'), ('always', 'RB'), ('the', 'DT'), ('future', 'NN'), ('.', '.'), ("'", "''"), (']', 'NN')]
[('[', 'IN'), ("'Europe", 'CD'), ('marks', 'NNS'), ('its', 'PRP$'), ('anniversary', 'NN'), (',', ','), ('it', 'PRP'), ('is', 'VBZ')]

我需要的是保留第二个值以“V”开头的每一行的元组

我尝试了很多方法，但我不知道如何：

 #df_my_string['clean_verbs_tokens']=filter((lambda x: x[1].startswith('V')),df_my_string[['verbs_tokens']])
#df_my_string['clean_verbs_tokens'] = df_my_string.verbs_tokens.apply(lambda x: str(x[0][1]).startswith('V'))
#df_my_string['clean_verbs_tokens'] = df_my_string.verbs_tokens.apply(lambda x: str(x[0][1]).startswith('V'))
#df_my_string['clean_verbs_tokens'] = [tup for tup in df_my_string['verbs_tokens'] if str(tup[0][1])=='V']
#df_my_string['clean_verbs_tokens'] = [item  for item in df_my_string['verbs_tokens'] if pd.Series(re.search('^V.*',item[0][1])).reset_index(drop=True).values]

预期输出：

verbs_tokens
[('was', 'VBD')]
[('is', 'VBZ')]

Answer 1

尝试：

df_my_string['clean_verbs_tokens'] = df_my_string["verbs_tokens"].apply(lambda x: [t for t in x if t[1].lower().startswith("v")])

>>> df_my_string['clean_verbs_tokens']
0    [(was, VBD)]
1     [(is, VBZ)]
Name: clean_verbs_tokens, dtype: object

Answer 2

# this is wrong because x is containing list of tuples
# so basically you are applying the condition only on 
# the first tuple
df_my_string['clean_verbs_tokens'] = df_my_string.verbs_tokens.apply(lambda x: str(x[0][1]).startswith('V'))

# try this
df_my_string['clean_verbs_tokens'] = df_my_string.verbs_tokens.apply(lambda x: [tup for tup in x if tup[1][0]=="V"])

Answer 3

这是一个解决方案：

df = pd.DataFrame(
    { 
       'Tuples' :  [ [('[', 'IN'), ("'Europe", 'CD'), ('marks', 'NNS'), ('its', 'PRP$'), ('anniversary', 'NN'), (',', ','), ('it', 'PRP'), ('is', 'VBZ')],
[('[', 'NNS'), ("'Europe", "''"), ('was', 'VBD'), ('always', 'RB'), ('the', 'DT'), ('future', 'NN'), ('.', '.'), ("'", "''"), (']', 'NN')] ]
    } )

定义一个函数来查找以任何字符开头的元组：

def find_char(tuples , char):
    start_with_char = []
    
    for tp in tuples: 
        if tp[1][:1]  == char:
                start_with_char.append(tp)
                
    return start_with_char

在您的数据框上应用该函数：

df['Tuples'].apply(lambda row: find_char(row ,'V')   )

结果：

0     [(is, VBZ)]
1    [(was, VBD)]

注意：此解决方案将为您提供包含字符

的元组列表

如何过滤 pandas 数据框中每一行的元组列表？

How to filter a list of tuples for every row in a pandas dataframe?

python

tuples

list

nltk

pandas