[ https://issues.jboss.org/browse/TEIID-2820?page=com.atlassian.jira.plugin... ] Ramesh Reddy commented on TEIID-2820: ------------------------------------- Couchbase seems to be much similar to MongoDB in terms of their document storage, type of query language supported. That could be lot more work, in magnitude of multiple weeks, depending upon how deep we want to integrate. Luckily we can draw upon the ideas of MERGE, EMBEDDED from MongoDB. For simple 1 to 1 bucket/table without nesting should be little less time consuming IMO. Either case one need to get very familiar with their N1QL language. INFER seems to be logical way to go about figuring out metadata. Otherwise you would need to write similar utility anyway. See Mongo, I had to do something similar. Good thing with Teiid is you can even avoid this for time being and define the schema at Teiid layer and develop mapping semantics to Couchbase, but would need the reverse at some point in the development. Yes, we need to follow the methodology similar to Simba driver above from [3]. We could have used Simba drivers, but that does not look like Open Source drivers, and I did not find any others either. Depending upon the community interest Teiid can be of good play in providing Open Source driver IMO. Also see [4] I would also poke the Couchbase community for any interest in opsn source JDBC/ODBC driver for this? May be someone with experience can help with our effort? [4] http://docs.couchbase.com/admin/admin/Views/views-translateSQL.html -- This message was sent by Atlassian JIRA (v7.2.3#72005)