https://typesense.org logo
Join Slack
Powered by
Hi there. What can I expect to happen with the sor...
# community-help
b
Hi there. What can I expect to happen with the sorting/ranking of documents when I use the option 
search_cutoff_ms
? I can see that the "found" results are smaller when I set it. Will the documents be ranked only within this subgroup? Does it mean that a perfect match that would be the first result in the same query without a cut off could be left out?
k
Yes your understanding is correct.
b
Thanks!
Is there a predictable mechanism to what will be prioritized and what will be left out? Like a default sorting in the index or something like that? I'm working with a dataset of 6M documents and some popular queries return as much as 1M documents. I'm thinking of setting 
search_cutoff_ms
in order to refrain these queries from overloading the server. Would that be the recommended path? It seems to work well, I'm just a little worried that I won't ever be displaying some documents that would match the query very well. Maybe it's not an issue as there's still a lot of documents being considered.
k
search_cutoff_ms
is a hard limit for how long the total query processing runs. Since we look up documents iteratively in terms of relaxing the typo / drop tokens thresholds, the good docs should be found ahead so there is a good chance that this will affect only matches with typos etc. But it's difficult to quantify exactly because there are tens of small steps that happen and each of them will be cognizant of this cutoff.
b
Got it. Thanks again!
Open in Slack
PreviousNext
Powered by