The following summary highlights what is new in Connect ETL 10.0 and later. Details of new features, platforms, and fixed problems can be found in their respective sections.
10.0.2
- Improved system security and stability by updating internal core components.
- Connect ETL is now fully supported on Windows Server 2025 and AIX 6.1 or later, providing you with more deployment options. For more information, see Prerequisites.
- New features have been added to Connect ETL.
- Improved EBCDIC code page support for DMXMFSRT/DMXMMSRT, now including euro currency variants for better compatibility and flexibility. For more information, see New functionality.
Fixed issues
- Jobs using Oracle OCI direct path—such as copy or ETL tasks—could fail during setup on Oracle servers running version 19.
- Running a task or job that concatenated a text constant containing an undefined environment variable could cause the job to abort.
For more information, see fixed issues for v10.0.2.
10.0.1
- Improved system security and stability by updating internal core components.
- New functionality has been added to Connect ETL.
- Added support for join functionalities
- Initial IPv6 support
For more information, see New functionality.
Fixed issues
- Resolved an issue where temporary tables created during a Connect ETL job were not automatically deleted after the job was created.
- Resolved an issue where records sharing the same key, but distributed across different groups, led to multiple partial counts. This was caused by a mismatch between the MapReduce sort key and the Hive table's partitioning scheme.
- Fixed a problem that introduced unexpected characters in the JDBC data connector.
For more information, see fixed issues for v10.0.1.
10.0
- Connect ETL now supports Java 11. For more information, see Prerequisites.
- New functionality has been added to Connect ETL.
- DMXMMSRT big-endian integer option
- Retrieval of Azure AD token
For more information, see New functionality.
Fixed issues
- Fixed the issue that caused the CNVRTOVR overflow warning to be deactivated and no longer issue alerts.
- Added target table information messages to tables and columns when some Snowflake data is skipped and errors are generated.
- Fixed an issue where records with out-of-range record length values caused conversion errors.
- Fixed the issue that prevented updates to the target table when data extraction from the source table failed.
- Fixed the issue where the Spark job failed to start due to a missing input path.
- Fixed the issue where output data was missing when the job aborted during execution of HIVE target database tables.
- Fixed an issue causing intermittent loss of records in long-running Hadoop MapReduce jobs during file joins, where some records from the left -side input were missing in the target output.
- Fixed the issue where Redshift write failure caused an I/O error on a connected file.
- Fixed an issue where large positive numbers were incorrectly converted to negative values during numeric processing, resulting in inaccurate averages.