Release Notes

Support for Spark 3.0.0.

Updated mongodb-driver-sync version to 4.0.4

Updated dependency to mongodb-driver-sync

Updated MongoDB Java Driver to 3.12.5.

Don't use Java SPI for the data source internally.

Fix BsonOrdering bug for strings of different lengths.

Ensures nullable fields or container types accept null values.

Added ReadConfig.batchSize property. For more information, see Input Configuration.

Renamed system property spark.mongodb.keep_alive_ms to mongodb.keep_alive_ms.

Added MongoDriverInformation to the default MongoClient.

Updated to latest Java driver (3.10.+)

Updated PartitionerHelper.matchQuery to no longer include $ne/$exists checks.

Added logging support for partitioners and their queries.

Added WriteConfig.extendedBsonTypes setting so users can disable extended BSON types when writing. For more information, see Output Configuration.

Added Java spi can now use short form: spark.read.format("mongo").

spark.read.format("mongo") can be used in place of spark.read.format("com.mongodb.spark.sql") and spark.read.format("com.mongodb.spark.sql.DefaultSource").

Support Spark 2.4.0. Updated Spark dependency to 2.4.0.

Ensure WriteConfig.ordered is applied to write operations.

Updated Mongo Java Driver to 3.9.0.

Added Scala 2.12 support.

Fixed MongoSpark.toDF() to use the provided MongoConnector.