[
https://issues.apache.org/jira/browse/SDAP-127?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Frank Greguska updated SDAP-127:
--------------------------------
Component/s: nexus
> ID should be unique in SOLR schema
> ----------------------------------
>
> Key: SDAP-127
> URL: https://issues.apache.org/jira/browse/SDAP-127
> Project: Apache Science Data Analytics Platform
> Issue Type: Improvement
> Components: nexus
> Reporter: Frank Greguska
> Priority: Major
>
> The "solr_id_s" field is currently the "uniqueKey" for the schema:
> https://github.com/apache/incubator-sdap-nexus/blob/107438af45b479348ffb75a667b276ee3c81f9da/data-access/config/schemas/solr/nexustiles/conf/managed-schema#L200
> This is fine but a lot of the algorithms depend on the simple "id" field for
> working with tiles (the id field is the same as solr_id_s but without the
> prefix used for document routing):
> https://github.com/apache/incubator-sdap-nexus/blob/107438af45b479348ffb75a667b276ee3c81f9da/data-access/config/schemas/solr/nexustiles/conf/managed-schema#L120
> If possible, the "id" field should also be marked as unique so that it is
> impossible to generate tiles with identical "id"s.
> This problem was found because of SLCP ice shelf data where 2 variables from
> the same granule were being ingested. The ID is generated from the granule
> name and section spec and an optional 'salt' value. In this case no salt was
> used (incorrectly) so the tiles were generated with identical "id"s but no
> error occurred because they had different dataset names which caused the
> "solr_id_s" field to be unique.
> Not sure if it is possible to have more than one unique field in a SOLR
> schema.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
