Teiid Couchbase Connector

Defining a Schema

The data stored in Couchbase are JSON documents, each document may contain:

  • key pair with no typed value
  • nested arrays
  • arrays of differently-typed elements
  • nested documents which contain above 3 scenarios

which does not follow the rules of data typing and structure that apply to traditional relational tables and columns. To map the documents to a relational form, the Teiid Couchbase connector will generates a database schema that maps the Couchbase data to a JDBC-compatible format. Simply, this be done in Teiid Connector’s MetadataProcessor load the Metadata.

Principles for mapping documents to JDBC-compatible Tables

  1. The keyspace/bucket be mapped as top table which contains all key/value pairs not include the nested arrays/documents, key be mapped to column name, the value type be mapped to column type
  2. The nested arrays/documents be mapped to different table with name: the parent table + underscore character + the key of nested arrays/documents
  3. Each Table has a PK column map the document id, the PK in top table is primary key.
  4. If a nested array has a nested array item, the array item item be treated as Object.

An Example of Schema Definition

There are documents Customer and Order under Bucket test:





Mapped table relations

Table relations

The above documents will map to 4 tables: test, test_CreditCard, test_Items, test_SavedAddresses.

Data Samples