Extract all matching substrings between two characters in the dataframe column-CodePudding

I have a dataframe column containing html.

I'm trying to extract everything between the list tags, (<li> and </li>) and return in a new column named 'output'.

I'd like to include the <li> and </li> tags too as part of the output.

Minimum Reproducible Example:

import pandas as pd

data = {
    'ID': ['1', '2'],
    'Description': ['blah blah <li>Point 1</li>blah blah<li>Point 2</li>blah blah blah', 'blah<li>Point1</li>blah<li>Point 2</li>']}

df = pd.DataFrame(data)
df['new'] = df.Description.apply(lambda st: st[st.find("<li>") 1:st.find("</li>")])
print(df)

Desired Output

  ID                                        Description         output
0  1  blah blah <li>Point 1</li>blah blah<li>Point 2...  <li>Point 1</l><li>Point 2</li>
1  2            blah<li>Point1</li>blah<li>Point 2</li>  <li>Point 1</l><li>Point 2</li>

What I've tried: Although there appears to be a lot of solutions around ('extracting substrings between two strings') there isn't anything that comes close.

For example, this (and many other result) only return the first instance Extract substring between two characters in pandas but I need to extract all instances.

Also I haven't seen any that will preserve the <li> tags.

CodePudding user response：

You can use look-behind assertion and look-ahead assertion to get your desired result. You can look into this amazing answer to understand more.

In your case, (?=<li>) will match the string followed by <li> while (?<=</li>) will match the string preceded by </li>.

result = df.Description.str.findall(r'(?=<li>).*?(?<=</li>)')

0    [<li>Point 1</li>, <li>Point 2</li>]
1     [<li>Point1</li>, <li>Point 2</li>]
Name: Description, dtype: object

Eventually, you assign it back to the dataframe.

df['output'] = result