DB Connector

[pkutrhari]

Hi - I am writing a DB connector (External Oracle DB) on the lines of the existing Modeshape JDBC Meta Connector (Modeshape 2.8.3).  Looking the Meta connector code, it looks like since the repo is in the cache, there is no Lucene involved, and hence ONLY path based queries can be done. Did I get it right? I tried doing a JCR-SQL2 query on the meta connector and no results were returned (although the data was in the repo at the time). Also, I am wondering how this is going to perform when huge sets of data is retrieved from the DB from various tables into the repository.

Has anyone else written a DB Connector and can talk about the performance and approaches etc.?

Thanks.

Praveen

[rhauch]

Since the content from the connector suddenly "appears" (and ModeShape thinks it's existing content), it is not indexed. You can rectify that by either having your connector publish events (e.g., signify the top-level node is new), or by reindexing the content using the ModeShape-specific Workspace interface that extends javax.jcr.Workspace. Either will work and will cause Lucene to index the content.

[pkutrhari]

Thanks Randall. I'll give that a shot.