[jira] [Commented] (SOLR-7510) UpdateProcessor to compute a murmur3 hash of a field at index time

["Hoss Man (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SOLR-7510?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14533141#comment-14533141 ] 

Hoss Man commented on SOLR-7510:
--------------------------------

Basic thinking i have...

* FieldMuttaingUpdateProcessor
* by default mutates no fields
* typical usage would be after CloneFieldUpdateProcessor
* looks at each field value given, and uses instanceof to pick the best method to call on the HashFunction
** if not a simple primitive, defaults to toString() then hash
** so for optimal hashing of numerics, users should put this after the appropriate Parse(Numer)UpdateProcessor
*** slightly cumbersome, but mainly targeted more for string fields anyway, since that's where pre-computing hte hash values is the most important

> UpdateProcessor to compute a murmur3 hash of a field at index time
> ------------------------------------------------------------------
>
>                 Key: SOLR-7510
>                 URL: https://issues.apache.org/jira/browse/SOLR-7510
>             Project: Solr
>          Issue Type: Sub-task
>            Reporter: Hoss Man
>
> SOLR-6968 is adding HyperLogLog support to stats component.  HLL accuracy depends on having good (long) hash values -- these can be computed at query time, but we should give users a simple option to compute them at index time for efficiency (especially with things like String hashing)



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org