Amazon Athena JDBC 3.x release notes

DataZone authentication – Added support for the DataZone authentication plugins DataZoneIdC and DataZoneIAM. For more information, see DataZone IdC Credentials Provider and DataZone IAM Credentials Provider.

Network timeout – The network timeout can now be set using the NetworkTimeoutMillis connection parameter. Previously it could be set only on the Connection object itself. For more information, see Network timeout.

S3 empty object handling – The driver now handles empty objects in the S3 fetcher instead of throwing an Amazon S3 Range Not Satisfiable exception.

Logging – The driver no longer logs the message Items requested for query execution [...], but subscription is cancelled after consuming query results.

Empty parameter strings – The driver now handles empty strings present in a connection parameter as if the parameter were not present. This resolves issues that occurred when some BI tools inadvertently passed empty strings that caused unintended authentication attempts.

Data type mapping – Improved the compliance with the JDBC spec by changing how the driver maps the tinyint, smallint, row, and struct data types to Java objects.

Amazon SDK version update – The Amazon SDK version used in the driver has been updated to 2.26.23.

Comments – Fixed an issue with line comments at the end of a statement.

Database listing – Fixed an issue in which listing databases could enter an infinite loop when the last page returned by the paginated ListDatabases API was empty.

JWT credentials provider – Added support for user-specified session durations. For more information, see Role session duration.

Thread pool – Created one ThreadPoolExecutor per connection for asynchronous tasks to avoid using the ForkJoin pool.

Credential providers – The proxy host is now parsed to get the scheme and host when the HTTP client is configured for external IdPs.

Default credentials provider – Ensured the default credentials provider can't be closed by client code.

getColumns – Fixed an ORDINAL_COLUMN column property issue in the DatabaseMetaData#getColumns method.

ResultSet – Added support for Infinity, -Infinity, and NaN to ResultSet. Fixed a discrepancy between the column type returned from catalog operations and the result set of a completed query.

Catalog operation performance – Performance has been improved for catalog operations that do not use wildcard characters.

Minimum polling interval change – The minimum polling interval default has been modified to reduce the number of API calls the driver makes to Athena. Query completions are still detected as soon as possible.

BI tool discoverability – The driver has been made more easily discoverable for business intelligence tools.

Data type mapping – Data type mapping to the Athena binary, array, and struct DDL data types has been improved.

Amazon SDK version – The Amazon SDK version used in the driver has been updated to 2.25.34.

Federated catalog table listings – Fixed an issue that caused federated catalogs to return an empty list of tables.

getSchemas – Fixed an issue that caused the JDBC DatabaseMetaData#getSchemas method to fetch databases only from the default catalog instead of from all catalogs.

getColumns – Fixed an issue that caused a null catalog to be returned when the JDBC DatabaseMetaData#getColumns method was called with a null catalog name.

Support added for Microsoft Active Directory Federation Services (AD FS) Windows Integrated Authentication and form-based authentication.

For backwards compatibility with version 2.x, the awsathena JDBC sub-protocol is now accepted but produces a deprecation warning. Use the athena JDBC sub-protocol instead.

AwsDataCatalog is now the default for the catalog parameter, and default is the default for the database parameter. These changes ensure that correct values for the current catalog and database are returned instead of null.

In conformance with the JDBC specification, IS_AUTOINCREMENT and IS_GENERATEDCOLUMN now return an empty string instead of NO.

The Athena int data type now maps to the same JDBC type as Athena integer instead of to other.

When the column metadata from Athena does not contain the optional precision and scale fields, the driver now returns zero for the corresponding values in a ResultSet column.

The Amazon SDK version has been updated to 2.21.39.

Fixed an issue with GetQueryResultsStream that caused an exception to occur when plain text results from Athena had a column count inconsistent with the column count in Athena result metadata.