It seems like the

_text_match

will always be higher if both search terms can be matched in one field vs. separate fields, no matter the

query_weight

. Is there any non-hacky way to get around this? Feels like it could be a parameter. Example: You have fields

title

and

type

and two documents

{"title": "Nike Shoerack", "type": "shoe_rack"}
{"title": "Nike Air Force 1", "type": "shoe"}

In this case I would like to be able to have

query_weight=1

for title and `query_weight=5`for the

type

to ensure that type is always the most important attribute. With my current setup, I get the shoerack first if I search for

nike shoe

@John Could you try setting

prioritize_exact_match=false

as a query param?

Does not change the results for me

This was the earlier behavior but there was a lot of unhappiness about this so we changed the approach to what we have now... Need to think about how to make both methods available via some flag.

Thanks Kishore, makes sense. It could be useful to be able to do both at the same time, e.g. with some “multiple_match_query_weight”, so that it’s not just on or off. Example:

multiple_match_query_weight=2

•

title

with

query_weight=1

•

description

with

query_weight=1

•

type

with

query_weight=5

Then

nike shoe

in title would be better than

nike

in title and

shoe

in description (because the text matches in title are multiplied by

2

or something like that), but not better than

nike

in title and

shoe

in type. Maybe it gets too complicated or doesn’t work in Typesense, just throwing the idea out there!