Release notes

Spark 2.4 Version 1.4.1

• Fix client re-initialization and fix async connection with Direct Mode
• Shade various libraries to support streaming in Azure Databricks via Maven co-ordinates

Spark 2.4 Version 1.4.0

• Add support for resource token based initialization
• Fix streaming error handling during partition split
• Upgrade Cosmos DB SDK to 2.4.1, BulkExecutor to 2.5.0 (with various bug fixes)

Spark 2.4 Version 1.3.5

• upgrade cosmos db sdk to 2.1.1
• upgrade jackson-databind to 2.9.8

Spark 2.4 Version 1.3.4

• Fix for sql query folding for precedence of AND and OR filters.
• Bump up the spark runtime dependency version to 2.4.0

Spark 2.3 Version 1.3.3

• Update the bulk executor dependency to be 2.4.1 version which includes fixes and update to dependencies
• Add support for database throughput collections.
• Fix spark sql query folding issue when there is a NOT operator by including brackets around each filter.
• Change default page size to 1000 when reading from Cosmos DB
• Fix a bug for reading a null array to read it as empty array