Hi all & @Kishore Nallan @Jason Bosco Need suggestion: Best Practices for Indexing Nested Fields in Typesense Schema I’m working with a dataset where the indexing requirements have evolved over time. Initially, I only needed to index a few fields, so I extracted and stored them as top-level fields to simplify searching. However, as the requirements have grown, I now need to index several nested fields. Here’s an example of my current data structure:

{
  "customer": {
    "name": "Customer Name"
  },
  "destination": {
    "appointment": {
      "start": {
        "date": "2024-01-01",
        "time": "12:00",
        "timezone": "America/New_York"
      }
    },
    "arrivalDateTime": {
      "date": "2024-01-02",
      "time": "14:00",
      "timezone": "America/New_York"
    },
    "departureDateTime": {
      "date": "2024-01-01",
      "time": "15:00",
      "timezone": "America/New_York"
    },
    "warehouse": {
      "address": {
        "city": "Destination City",
        "stateCode": "ST"
      }
    }
  },
  "origin": {
    "appointment": {
      "start": {
        "date": "2024-01-01",
        "time": "08:00",
        "timezone": "America/New_York"
      }
    },
    "arrivalDateTime": {
      "date": "2024-01-01",
      "time": "10:00",
      "timezone": "America/New_York"
    },
    "departureDateTime": {
      "date": "2024-01-01",
      "time": "11:00",
      "timezone": "America/New_York"
    },
    "warehouse": {
      "address": {
        "city": "Origin City",
        "stateCode": "ST"
      }
    }
  },
  "drivers": [
    {
      "phoneNumber": "1234567890"
    }
  ],
  "dispatchers": [
    {
      "email": "<mailto:dispatcher@example.com|dispatcher@example.com>"
    }
  ]
}

Fields I now need to index include: •

destination.appointment.start.date

•

destination.arrivalDateTime.date

•

destination.departureDateTime.date

•

destination.warehouse.address.city

•

destination.warehouse.address.stateCode

•

origin.appointment.start.date

•

origin.arrivalDateTime.date

•

origin.departureDateTime.date

•

origin.warehouse.address.city

•

origin.warehouse.address.stateCode

•

customer.name

•

drivers.phoneNumber

•

dispatchers.email

My Questions: 1. Flattening vs. Nested Indexing: Should I extract all these fields to top-level keys for indexing, or is it better to keep them nested? ◦ Are there performance implications or best practices to consider? 2. Search Optimization: Would indexing these deeply nested fields directly impact search performance compared to flat fields? 3. Future Scalability: If more nested fields need to be indexed later, how should I design my schema to accommodate these changes efficiently? Any advice or suggestions for structuring my data and schema to handle these requirements effectively would be greatly appreciated. Thank you!

The performance impact of nested records is minimal. During indexing, there's a small overhead from flattening nested JSON fields into dot notation (e.g.,

destination.warehouse.address.city

). During searching, there's a minor overhead from deleting these flattened fields before returning results, but it shouldn't even be noticeable

Great, Thanks