Dynamic regex in dataframe

Question

Having a dataframe as follows:

df= pd.DataFrame({'category':['Fishing','Refrigeration','store'],'synonyms_text':['seafood','foodlocker',' food']})

And a list as follows:

list_desc=['FOOD', 'GROWERS', 'INTERNATIONAL']

How can I iterate over the list_desc to create a dynamic regular expression to be used in the dataframe?

for word in list_desc:
    print(word.lower())
    df_tmp= df.loc[df['synonyms_text'].str.contains(r'\bfood\b')]

Where food has to be substituted by word variable.

Thanks

There are lots of options, and watch out for caveats. See [the answer](https://stackoverflow.com/a/29996092/3832970). — Wiktor Stribiżew, Oct 31 '19 at 00:53

Rithin Chalumuri · Answer 1 · 2019-10-31T01:02:05.257

0

You can just construct your regex dynamically using format() like in r'\b{0}\b'.format(word)

Example:

for word in list_desc:
    df_tmp= df.loc[df['synonyms_text'].str.contains(r'\b{0}\b'.format(re.escape(word.lower())))]

edited Oct 31 '19 at 01:02

answered Oct 31 '19 at 00:39

Rithin Chalumuri

Perfect, it worked. Thanks – John Barton Oct 31 '19 at 00:44
@JuanPerez, glad that it's solved. And if my answer is helpful, could you please help accept it? Thanks :) – Rithin Chalumuri Oct 31 '19 at 00:49
1

It won't work once the `word` contains special chars. Word boundaries will stop working once the special char appears at the start/end of the word. – Wiktor Stribiżew Oct 31 '19 at 00:53
1

better to use `f-string` for stringformatting – Erfan Oct 31 '19 at 00:54
@WiktorStribiżew, will add re.escape to the solution. – Rithin Chalumuri Oct 31 '19 at 01:01

1 Answers1