[
https://issues.apache.org/jira/browse/SOLR-6965?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14277390#comment-14277390
]
Grant Ingersoll commented on SOLR-6965:
---------------------------------------
One of the main use cases for me is that CSV is often only single valued, so
perhaps we would guess better there. See also SOLR-6979
> Consider passing MIME-type info into field guessing capabilities
> ----------------------------------------------------------------
>
> Key: SOLR-6965
> URL: https://issues.apache.org/jira/browse/SOLR-6965
> Project: Solr
> Issue Type: Improvement
> Reporter: Grant Ingersoll
>
> In digging in on data-driven/field guessing/schemaless a bit more, my gut
> instinct after staring at lots of different file types is that we should, if
> possible, pass MIME type info through to the guessing mechanism so that we
> can potentially make different choices for different types. For instance,
> CSV is much more structured and can likely be smarter about data than XML or
> PDF. Same goes for JSON.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
