Custom dataset preparation

[azz azz]

I am currently working on creating a custom Question answer dataset and I would like to get some tips apart from what's written on the paper.

I faced one issue where the answer of a particular question is not continuous, it is the headings of different paragraphs. How do I go about solving this?

For example

Paragraph:

(a) Nudity: 

         No content that is prohibited by law at the time being in force can be published or transmitted.

(b) Sex: 

        No content that is prohibited by law at the time being in force can be published or transmitted. The non-explicit (implicit) to explicit depiction of sexual behaviour. 

(c) Violence: 

         Classification decisions shall take account of the degree and nature of violence in a work.  

Question: What are the types of behavioural issues that can be portrayed in films?

Answer: Nudity, Sex, Violence

Nudity sex and violence are different parts of the paragraph.

Your help would be highly appreciated.