Python: Regex – Matching Foreign Characters/Unicode Letters
In the world of regular expressions, matching characters outside of the usual Latin character set can be a challenge. Read on to find out how author Mark Needham tackled this issue in Python.
Join the DZone community and get the full member experience.
Join For FreeI’ve been back in the land of screen scraping this week extracting data from the Game of Thrones wiki and needed to write a regular expression to pull out characters and actors.
Here are some examples of the format of the data:
|
So the pattern is:
<actor> as <character>
Optionally followed by some other text that we’re not interested in.
The output I want to get is:
|
I started using the ‘split’ command on the word ‘as’ but that broke down when I realized some of the characters had the letters ‘as’ in the middle of their name. So, regex it is!
This was my first attempt:
|
|
It works for 4 of the 5 scenarios but not for Filip Lozić. The ‘ć’ character causes the issue so we need to be able to match foreign characters which the current charset I defined in the regex doesn’t capture.
I came across this Stack Overflow post which said that in some regex libraries you can use ‘\p{L}’ to match all letters. I gave that a try:
|
And then re-ran the script:
|
Hmmm, not sure if I did it wrong or if that isn’t available in Python. I’ll assume the latter but feel free to correct me in the comments and I’ll update the post.
I went search again and found this post which suggested another approach:
You can construct a new character class:
[^\W\d_]
instead of \w. Translated into English, it means “Any character that is not a non-alphanumeric character ([^\W] is the same as \w), but that is also not a digit and not an underscore”.
Let’s try plugging that in:
|
|
So that’s fixed Filip but now Daniel Naprous is being incorrectly parsed.
For Attempt #4 I decided to try excluding what I don’t want instead:
|
|
That does the job but has exposed my lack of regex skills. If you know a better way let me know in the comments.
Published at DZone with permission of Mark Needham, DZone MVB. See the original article here.
Opinions expressed by DZone contributors are their own.
Comments