CODEWITHSUNDEEP
solr
stinger
stinger

Reputation: 4138

Create index of nested documents in SOLR

How should I import nested entities from DB to Solr index? For some reasons i don't want to flatten documents into single one. What should i write in schema.xml and data-config.xml ? I'm using Solr 4.10.

Upvotes: 3

Views: 863

Answers (2)

Josh Edwards
Josh Edwards

Reputation: 908

Since you can't use the DataImportHandler, you could write custom code to do this. I'd recommend using SolrJ to load childDocuments. To handle childDocuments, first you have to create all of your required fields (for all of your different record types) in your schema.xml (or use dynamic fields). From there, you can create a SolrInputDocument for the parent, and a SolrInputDocument for the child, and then call addChildDocument(doc) on the parent SolrInputDocument to add the child to it.

I'd also recommend creating a field that can indicate what level you're at - something like "content_type" that you fill in with "parent" or "root," or whatever works for you. Then, once you've loaded the records, you can use the Block/Join Queries to search hierarchically. Be aware that doing this will create an entry for each record, though, and if you do a q=: query, you'll get all of your records intermixed with each other.

Upvotes: 2

MatsLindh
MatsLindh

Reputation: 52792

The currently distributed version of the DataImportHandler does not support nested documents (or BlockJoins as they're called in Solr/Lucene).

There is however a patch available that you can try out - be sure to follow the discussion on JIRA (SOLR-5147) about how to use it and where it goes in the future.

Upvotes: 2

Related Questions