What's New in RapidMiner Studio 9.10.6?
Released: Mar 24th, 2022
RapidMiner Studio 9.10.2 - 9.10.5 have been skipped and were not released.
The following describes the bug fixes in RapidMiner Studio 9.10.6:
- Fixed a memory & file leak when using large numbers of repeated JDBC connections
- Visualizations: Added options to customize Wordcloud word orientations
- Visualizations: Added Jamaica to the map collection
- Updated postgres JDBC driver to version 42.3.2
- Added skip inaccessible parameter for Loop Files to skip inaccessible files/directories, instead of a silent failure. If unchecked, the operator does not loop at all and will throw a proper error.
- Stopping Loop Files is now always possible in a timely manner, even if you selected a directory with millions of files.
- Updated H2 DB library due to security advisory
- Added new parameter fitting error handling to the ARIMA Trainer operator.
- In case of a fitting error during training, either a proper error is thrown or a fallback Default Forecast Model is provided.
- Removed meta data warning for number of parameters is too large for the ARIMA Trainer operator.
- Added new option to Amazon S3 connections that allows for much more flexible authentication schemes, like credential profiles and IAM roles.
- Fixed character corruption issue with Read Database and Execute SQL when reading a query via a file from disk on certain operating systems
- Fixed a memory leak when using database connections
- Fixed a general file leak when using connections
- Fixed a problem when creating dynamically suffixed attributes through the AttributeFactory in parallel
- Fixed side effects for models when executing in parallel
- Fixed an issue in projects that could sometimes cause Execute Process or Retrieve operators within parallel loops or similar setups to fail with an error message like "Cannot retrieve 'entry', it does not exist"
- Fixed an issue that could sometimes cause Execute Process operators within parallel loops or similar setups to fail with a error messages like "Cannot connect to the RapidMiner AI Hub repository '_LOCAL'" when running on an AI Hub legacy repository
- Fixed a wrong error, which was thrown during Apply Forecast when a Multiply operator was used on the Holt-Winters model
- Fixed calculation errors for Holt-Winters models with additive seasonality