Use a flat or nested record structure for indexing news website?

Open Q&A
1

Hello, I’m looking for advice on how to index our records. The tldr is should we structure our records flat or nested?

Context: we have a WordPress website of news articles, we want users to be able to search for posts (articles), pages, authors and tags and categories. So a search result could be a author or a post or a category.

  • Should we structure our records flat but with different ‘record types’, one for author, one for post, etc.?
  • Should all records have the same properties? For eg, if a post record has a excerpt property should a tag and author record also have excerpt even though its not relevant to those types of records?

Eg of flat record structure in an index.

// Post
{
    "type": "post",
    "title": "",
    "excerpt": "",
    "url": "",
    "image": "",
    "content": "",
    "date": ""
}

// Page
{
    "type": "page",
    "title": "",
    "excerpt": "",
    "url": "",
    "image": "",
    "content": "",
    "date": ""
}

// Author
{
    "type": "author",
    "name": "",
    "url": "",
    "image": ""
}

// Term
{
    "type": "category|tag",
    "name": "",
    "url": "",
    "image": ""
}

Eg of nested


// Post or page
{
    "type": "post|page",
    "title": "",
    "excerpt": "",
    "url": "",
    "image": "",
    "content": "",
    "date": "",
    "author": {
        "name": "",
        "url": "",
        "image": ""
    }
    "tags": [
        {
            "type": "category|tag",
            "name": "",
            "url": "",
            "image": ""
        },
        ...
    ]
}