Feature #6346
closed
Make # READ events available in SOLR index
Added by ben leinfelder over 10 years ago.
Updated over 10 years ago.
Estimated time:
(Total: 0.00 h)
Description
I'd suggest using the dynamic fields we have at our disposal: https://cwiki.apache.org/confluence/display/solr/Dynamic+Fields
Something like "read_count_i" for the dynamicField name.
We need to add a channel for getting name/value pairs into the index that is not based on the document content, but any arbitrary code that wants to include values in the index for a given document.
- Description updated (diff)
- Assignee set to Peter Slaughter
- Status changed from New to In Progress
- Assignee changed from Peter Slaughter to ben leinfelder
I committed my first attempt at this. It should probably be tested a bit more to ensure that indexing events do not unintentionally overwrite the additional field[s] added to the index that are not derived from the document or SystemMetadata. I intend to use this same mechanism to augment the index with semantic fields.
- Status changed from In Progress to Resolved
Note that this uses our own copy of the solr schema.xml instead of the version offered by the DataONE library. I think this is okay and I have committed the same changes to the trunk of the d1 library so that the multi-valued dynamic fields will be available for our use.
While this is not strictly required for the READ events (there is only one value) the mechanism that indexes those events does generally allow for multivalued dynamic fields being added to the index. So if the schema does not have those declared, it will fail to add those updates to the index.
Also available in: Atom
PDF