修剪数据框中的特定单词

问题描述

我有一个带有一些三元组（以及更多 ngram）的 df，我想检查句子是否以特定单词列表开头或结尾，并将它们从我的 df 中删除。例如：

import pandas as pd
df = pd.DataFrame({'Trigrams+': ['because of tuna','to your family','pay to you','give you in','happy birthday to you'],'Count': [10,9,8,7,5]})

list_remove = ['of','in','to','a']

print(df)

    Trigrams+            Count
0   because of tuna       10
1   to your family         9
2   pay to you             8
3   give you in            7
4   happy birthday to you  5

我尝试使用 strip 但在上面的示例中，第一行将返回 因为 tun

输出应该是这样的：

list_remove = ['of','a']

    Trigrams+             Count
0   because of tuna        10
1   pay to you              8
2   happy birthday to you   5

有人可以帮我吗？提前致谢！

解决方法

试试：

list_remove = ["of","in","to","a"]

tmp = df["Trigrams+"].str.split()

df = df[~(tmp.str[0].isin(list_remove) | tmp.str[-1].isin(list_remove))]
print(df)

打印：

               Trigrams+  Count
0        because of tuna     10
2             pay to you      8
4  happy birthday to you      5

,

你可以试试这样的：

import numpy as np

def func(x):
  y = x.split()[0]
  z = x.split()[-1]
  if (y in list_remove) or (z in list_remove):
     return np.nan
  return x

df['Trigrams+'] = df['Trigrams+'].apply(lambda x:func(x))
df = df.dropna().reset_index(drop=True)