[
https://issues.apache.org/jira/browse/SOLR-8092?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Joel Bernstein updated SOLR-8092:
---------------------------------
Description:
Currently the facet DocSet generated during the first phase of the distributed
search is not cached. This means that during the facet refinement phase the
DocSet needs to be recreated. This involves re-issuing the entire query and
re-building the DocSet. In some use cases that can be a high price to pay to
calculate refinements.
This ticket aims to find a way to cache the DocSet generated from the first
phase so it can be re-used in the second phase.
We may also want to add a way to turn off refinements as some use cases might
want to trade off performance for accuracy.
was:
Currently the facet DocSet generated during the first phase of the distributed
search is not cached. This means that during the facet refinement phase the
DocSet needs to be recreated. This involves re-issuing the entire query and
re-building the DocSet. In some use cases that can be a high price to pay to
calculate refinements.
This ticket aims to find a way to cache the DocSet generated from the first
phase so it can be re-used in the second phase.
We may also want to add a way to turn off refinements as some use case might
want to trade off performance for accuracy.
> The same faceting DocSet is generated in both the first and second phase of
> distributed search if there are facet refinements.
> ------------------------------------------------------------------------------------------------------------------------------
>
> Key: SOLR-8092
> URL: https://issues.apache.org/jira/browse/SOLR-8092
> Project: Solr
> Issue Type: Improvement
> Reporter: Joel Bernstein
>
> Currently the facet DocSet generated during the first phase of the
> distributed search is not cached. This means that during the facet refinement
> phase the DocSet needs to be recreated. This involves re-issuing the entire
> query and re-building the DocSet. In some use cases that can be a high price
> to pay to calculate refinements.
> This ticket aims to find a way to cache the DocSet generated from the first
> phase so it can be re-used in the second phase.
> We may also want to add a way to turn off refinements as some use cases might
> want to trade off performance for accuracy.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
