CODEWITHSUNDEEP
ruby-on-railsrubylucenesolrsunspot
clyfe
clyfe

Reputation: 23770

Solr partial document index update

I'm using Solr and Solr:Cell plugin to index and search rich text documents and metadata.

DEFINITION: solr_document = tuple(rich_text_document, metadata1, metadata2)

I want to reindex some solr_documents when metadata changes, but only the parts in the solr_document that chaged, not the whole solr_documnt, because parsing and extracting text from rich text documents is computing expensive and pointless since the rich text document was not modified.

Does Solr support partial document index updates?

Aditional:
I'm using Solr via sunspot in a Rails application. One of the main sunspot developers says here that:

Solr does not support the concept of partial updates -- under the hood, updating a document actually consists of removing it from the index and then re-adding it. So, Sunspot does have to construct the full document each time anything changes; it's an unfortunate limitation from a performance standpoint, but it's pretty fundamental to the way Solr and Lucene work.

Is there anything that can be done, maybe Solr:Cell allows something?

Upvotes: 3

Views: 1583

Answers (1)

Mauricio Scheffer
Mauricio Scheffer

Reputation: 99750

Does Solr support partial document index updates?

Nope. Check out the FAQ.

Is there anything that can be done?

Yes, IIRC there was an issue in the project JIRA about it. Look it up, ask what's missing, contribute to the effort of implementing it.

Upvotes: 3

Related Questions