[teiid-issues] [JBoss JIRA] (TEIID-2820) Support Couchbase as a resource

[ https://issues.jboss.org/browse/TEIID-2820?page=com.atlassian.jira.plugin... ] Kylin Soong commented on TEIID-2820: ------------------------------------ I would like to write a summary and difficults that I have met so far, after some investigation of Couchbase documents and samples. * The big difficults that block me so far is that it's hard to get the metadata of a specific document. In a Couchbase *Cluster*(same as a Database in RDBMS), there are lots of *Bucket*(same as Keyspace in NoSQL, or Table in RDBMS), each *Bucket* can contains lots of *Document*. These documents are JSON, use _a self-describing format, flexible Data Model, dynamic schemas_, and each document can have different attributes, can not get the attributes' name and data type in my investigation. * Community version can test limit features, Couchbase has 2 version: Community version and Enterprise version, the Enterprise version seems have full rights to access all features and a better documents, [1] hints there are SQL can infer the metadata(a JSON schema as [2]) of documents in a keyspace, for example the structure of documents, data types of various attributes, sample values, and so on, but The INFER statement is only supported in the Enterprise Edition. * Another approach to integrate with JDBC way is to define the metadata, mapping the documents in Bucket/keyspace to several Tables, including define data columns/data types, etc, An example is Simba JDBC Drivers, [3] is how schema definition in Simba. Any suggests and advices? [1] https://developer.couchbase.com/documentation/server/4.6/n1ql/n1ql-langua... [2] http://json-schema.org/documentation.html [3] http://cdn.simba.com/products/Couchbase/doc/JDBC_InstallGuide/content/jdb... -- This message was sent by Atlassian JIRA (v7.2.3#72005)