CreativePro Forum

Return to Member Forum

• Author
  Posts
• • November 18, 2016 at 2:38 am #89885
    Thijmen Van Brunschot

    Member

    Hello everyone,

    For an art project I am trying to scrape through large amounts of text (about 4000 A4) and want to scrape the questions in the text.
    To explain myself further see the example below.

    Here is just any text with a period. This is the question I want to scrape? Here is just any text with a period.

    Now I would like just to scrape the following part: “This is the question I want to scrape?”
    Does anyone have any idea on how this can be achieved?

    I tried \>\? but that only finds the question mark itself.

  • November 18, 2016 at 7:17 am #89887
    David Blatner

    Keymaster

    There is probably a more elegant way to do this but this is a start:
    (\w||'|,|;|~_|~=)+?\?
    That means look for a string of characters that could be “a letter or number, a space, an apostrophe, a comma, a semicolon, an em dash, or an en dash” and that ends with a question mark. (The vertical bar means “or”)

  • November 18, 2016 at 10:51 am #89900
    Aaron Troia

    Participant

    David is on the right track, I would do it a little differently with a positive lookbehind or a Keep (\K) to look for (but not capture) the punctuation and space before, such as:

    (?<=\. ).+?\?
    or
    \. \K.+?\?

    they both do the same thing, just in a different way. They both look for a period space (both zero width so as not to be included or affected in your replace) on one end and a question mark on the other end, and, of course, everything in between.

    you can either then group the .+? like (.+?) as a capture group or if you can just use it as is and in your replace field use $0 which basically matches everything that was found.

    Let me know if you have any questions.

    Aaron

  • November 18, 2016 at 11:22 am #89901
    David Blatner

    Keymaster

    Thanks, Aaron! Good to hear from you. Those expressions are great. But that won’t work if the question is the first sentence in the paragraph, right? Or if the previous sentence ends with an exclamation point or some other non-period punctuation.

  • November 18, 2016 at 12:00 pm #89902
    Aaron Troia

    Participant

    Hey David! Yes you’re right, I should’ve taken that into account, I have modified the version using Keep (\K) to account for more instance, and you could add more punctuation into the noncapture group at the beginning ((?:^|\.|\!|\?)) to account for more. There might be a better way, I know this isnt perfect, but it should catch most instances.

    (?:^|\.|\!|\?) ?\K.+?\?

  • November 20, 2016 at 6:55 am #89913
    Thijmen Van Brunschot

    Member

    Thank you David and Aaron, that helped a lot. I got what I was looking for!

• Author
  Posts