Another question, is there any speed degradation when values are in a list vs flattened? I am trying to think of how the above table would look as a single collection. If a brand can have up to 3 names, and unlimited addresses (think something like Target which have tens of thousands) should I have my collection be 1.

names: list[str], state: str, city: str, brand_id: str

2.

name: str, state: str, city: str, brand_id: str

3.

names: list[str], addresses: list[{state: str, city: str}], brand_id: str

If someone searches for

q=Target,NY,New York query_by=names,state,city

is there any speed increase/decrease between the above options?

Array fields may have negligible performance hits as opposed to non-array ones, but you shouldn't notice any difference

So just to make sure im understanding things correctly, you are saying the above 3 options should have negligible difference in speed/throughput? Would there be differences in terms of scoring rows? My guess is if I did

q=Targ NY

that the row for Target with 10k NY locations would match higher than a row for Targ with 1 NY locations? Due to the large difference in number of matched tokens on the row?

More matches in an array field will have a higher score of course (multiple Targ NY values in the

names

field)