I'd like to find text in one field of a pandas dataframe ("text") based on another field ("words") of it.
#import re
import pandas as pd
df = pd.DataFrame([['I like apple pie','apple'],['Nice banana and lemon','banana|lemon']], columns=['text','words'])
df['text'] = df['text'].str.replace(r''+df['words'].str, '*'+group(0)+'*')
df
I'd like to mark the found words with *.
How can I do that?
The desired output is:
I like *apple* pie
Nice *banana* and *lemon*
You could capture the word from words and use backreference in the substitution to wrap it in *:
import re
import pandas as pd
df = pd.DataFrame([['I like apple pie','apple'],['Nice banana and lemon','banana|lemon']], columns=['text','words'])
df['text'] = df['text'].replace(r'('+df['words']+')', r'*\1*', regex=True)
print(df)
Prints:
text words
0 I like *apple* pie apple
1 Nice *banana* and *lemon* banana|lemon
IIUC using (?i) is same as re.I
df.text.replace(regex=r'(?i)'+ df.words,value="*")
Out[131]:
0 I like * pie
1 Nice * and *
Name: text, dtype: object
Since you update the question
df.words=df.words.str.split('|')
s=df.words.apply(pd.Series).stack()
df.text.replace(dict(zip(s,'*'+s+'*')),regex=True)
Out[139]:
0 I like *apple* pie
1 Nice *banana* and *lemon*
Name: text, dtype: object
