Python Pandas.apply在spacy doc列上返回none值

我在我的“sp500news3”上运行以下命令，它返回一个None值

def extract_ticker(title):
    for word in title:
        if word in constituents['Symbol']:
            return word

sp500news3['tickers'] = sp500news3['title'].apply(extract_ticker)

#sp500news3 sample:



  index date_publish    title   tickers
0   79944   2007-01-29 19:08:35 (MSFT, Vista, corporate, sales, go, very, well) None
1   181781  2007-12-14 19:39:06 (WMB, No, Anglican, consensus, on, Episcopal, Church)   None
2   213175  2008-01-22 11:17:19 (CSX, quarterly, profit, rises) None
3   93554   2008-01-22 18:52:56 (C, says, 30, bln, capital, helps, exceed, target)  None

成分['Symbol']：样本

0      TWX  
1      C  
2      MSFT  
3      WMB ...

从以下位置复制spacy文档：

constituents =  pd.DataFrame({"Symbol":["TWX","C","MSFT","WMB"]})

sp500news3 = pd.DataFrame({"title":["MSFT Vista corporate sales go very well","WMB No Anglican consensus on Episcopal Church","CSX quarterly profit rises",'C says 30 bln capital helps exceed target','TWX plans cable spinoff']})

import spacy

nlp = spacy.load('en_core_web_sm')

sp500news3['title'] = sp500news3['title'].apply(nlp)

---

您必须使用

word.text

，因为当它迭代：

---

以你的例子：

In [11]: sp500news3['title'].apply(extract_ticker)
Out[11]:
0    MSFT
1     WMB
2    None
3       C
4     TWX
Name: title, dtype: object

---

你期望发生什么？

成分中的字符串似乎都不在您的标题中。标题是什么数据类型？它是字符串还是元组？据推测，MSFT在成分中，这是第一个成分的预期结果？成分比样本长-包含所有sp500标记-期望从每个标题中提取标记并添加到
sp500news3
DF中的标记列标题的数据类型是
spacy.tokens.doc.doc
我想不是这样的“不为字符串实现
\uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu
：”但熊猫不像为字符串实现的那样对其进行特殊处理。这仍然不返回任何值，检查了
components
数据类型，即
str
@W.R components是字符串，但title不是（您说它是Doc）。请分享一个我们可以在其上复制的示例。在发送到spacy doc之前已添加df，spacy doc的代码已添加
成分
df
In [11]: sp500news3['title'].apply(extract_ticker)
Out[11]:
0    MSFT
1     WMB
2    None
3       C
4     TWX
Name: title, dtype: object