Question

如果列中的条目包含列表中的单词，我想在我的python pandas数据帧中添加一个标志

对于我们可以使用的任何separete行

any(word in train['a'][0] for word in words)

我试图制作一个模式

import pandas as pd
import numpy as np
words=['photos','pictures', ' pics ', 'pix', 'image']
pattern = '|'.join(words)

train=pd.DataFrame()
train['a']=words

我尝试过使用contains但它没有获得模式

def emb_col_1(tr, te, col, pat, suf):
    tr["0_"+col+suf]=0
    tr.loc[tr[col].str.contains(pat), "0_"+col+suf] =1
    #tr.loc[tr[col].str.count(pat)>0, "0_"+col+suf] =1
    #tr.loc[(word in tr[col].str for word in pat), "0_"+col+suf] =1
    #tr["0_"+col+suf] = np.where(tr[col].str.contains(pat, case=False, na=False), 1, 0)
    #tr["0_"+col+suf] = np.where(any(word in train[col] for word in pat), 1, 0)


emb_col_1(train, test, 'a', words, '_p')
emb_col_1(train, test, 'a', pattern, '_p')

提前谢谢

Answer 1

我相信你需要：

words=['photos','pictures', ' pics ', 'pix', 'image']
#remeove trailining whitespaces by strip
pattern = '|'.join([x.strip() for x in words])

train=pd.DataFrame()
#added more values for test
train['a']=words + ['a','pics sss']
print (train)

#remove unused te
def emb_col_1(tr, col, pat, suf):
    #convert True and Falses to 1 and 0 by astype(int)
    tr["0_"+col+suf] = tr[col].str.contains(pat, case=False, na=False).astype(int)
    #return DataFrame
    return tr


df1 = emb_col_1(train, 'a', pattern, '_p')
print (df1)
          a  0_a_p
0    photos      1
1  pictures      1
2     pics       1
3       pix      1
4     image      1
5         a      0
6  pics sss      1

编辑：

words=['photos',' pics ', 'pix', 'image']
#remeove trailining whitespaces by strip
pattern = '|'.join([r'\b{}\b'.format(x.strip()) for x in words])

train=pd.DataFrame()
#added more values for test
train['a']=words + ['a','pics sss', 'pictures']
print (train)
          a
0    photos
1     pics 
2       pix
3     image
4         a
5  pics sss
6  pictures

#remove unused te
def emb_col_1(tr, col, pat, suf):
    #convert True and Falses to 1 and 0 by astype(int)
    tr["0_"+col+suf] = tr[col].str.contains(pat, case=False, na=False).astype(int)
    #return DataFrame
    return tr


df1 = emb_col_1(train, 'a', pattern, '_p')
print (df1)
          a  0_a_p
0    photos      1
1     pics       1
2       pix      1
3     image      1
4         a      0
5  pics sss      1
6  pictures      0

创建包含列表中的单词的列标志

1 个答案: