Although Apache Spark SQL does not define a Clob data type, the driver allows you to retrieve and update long data, specifically LONGVARCHAR data, using JDBC methods designed for Clobs. To do this, the StringDescribeType property must be set to longvarchar.

When using these methods to update long data as Clobs, the updates are made to the local copy of the data contained in the Clob object.

Note: The driver does not support retrieving and updating long data using JDBC methods designed for Blobs.

Retrieving and updating long data using JDBC methods designed for Clobs provides some of the same benefits as retrieving and updating Clobs, such as:

  • Provides random access to data
  • Allows searching for patterns in the data, such as retrieving long data that begins with a specific character string

To provide these benefits normally associated with Clobs, data must be cached. Because data is cached, your application will incur a performance penalty.