This message was deleted.
# community-help
s
This message was deleted.
k
Try with
max_candidates: 10
-- by default Typesense only looks at 4 candidates by default. You could also try setting
popularity
as the
default_sorting_field
in collection schema. This will make Typesense pick popular fields first, instead of going by token frequency while picking max candidates.
๐Ÿ‘€ 1
However, on second thoughts, given that the expected results are returned (i.e. keyword is identified) but it's not ordered, maybe there is another issue here. What happens when you remove the buckets: 10?
m
The same
โ€œconadโ€ never shows up until the 3rd character
k
Got it, then it's probably the max candidates issue.
If that doesn't work, I might need to see the json response returned. For the "c" query. You can DM me and if you wish mask out sensitive content.
m
How do max candidates interact with grouping?
k
Max candidates is just a list of tokens that Typesense will search for when there are too many candidates, for e.g. for shorter prefixes.
So doesn't directly interact with grouping.
m
ok, the max candidates did the job
thx a lot
๐Ÿ‘ 1
I had to increase it to 50
k
Default sorting field didn't help?
m
I already had it properly configured
is there any performance concern with max candidates to 50?
k
Depends on the size of your dataset.
m
45k
k
Should be fine.
๐Ÿ™Œ 1