[ 
https://issues.apache.org/jira/browse/SOLR-9267?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Ere Maijala updated SOLR-9267:
------------------------------
    Attachment: SOLR-9267.patch

This patch fixes the handling of qf so that fields and any boosts are always 
extracted. It also fixes the filteredDocument creation so that IndexableField 
type is not directly cast to a string as that would include strings like 
"indexed" and "stored" and throw the results off if those are included in the 
indexed records.

> Cloud MLT field boost not working
> ---------------------------------
>
>                 Key: SOLR-9267
>                 URL: https://issues.apache.org/jira/browse/SOLR-9267
>             Project: Solr
>          Issue Type: Bug
>      Security Level: Public(Default Security Level. Issues are Public) 
>          Components: MoreLikeThis
>    Affects Versions: 5.0, 5.1, 5.2, 5.2.1, 5.3, 5.3.1, 5.3.2, 5.4, 5.4.1, 
> 5.5, 5.5.1, 5.5.2, 5.5.3, 5.6, 6.0, 6.0.1, 6.0.2, 6.1, 6.1.1, 6.2
>            Reporter: Brian Feldman
>         Attachments: SOLR-9267.patch
>
>
> When boosting by field "fieldname otherFieldName^4.0" the boost is not 
> stripped from the field name when adding to fieldNames ArrayList.  So on line 
> 133 of CloudMLTQParser when adding field content to the filteredDocument the 
> field is not found (incorrectly trying to find 'otherFieldName^4.0').
> The easiest but perhaps hackiest solution is to overwrite qf:
> {code}
> if (localParams.get("boost") != null) {
>       mlt.setBoost(localParams.getBool("boost"));
>       boostFields = SolrPluginUtils.parseFieldBoosts(qf);
>       qf = boostFields.keySet().toArray(qf);
> }
> {code}



--
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

Reply via email to