What's New in RapidMiner Studio 9.1.0?
Released: December 13th, 2018
The following describes the bug fixes in RapidMiner Studio 9.1.0:
- The Aggregate Operator got the percentile function where the percentile can be changed in the aggregation attributes functions list. It is possible to use an integer like 75 or a floating point value like 80.5 here. It is of course also possible to use a macro here.
- Split the setting to keep operators connected upon disabling or deleting them into these settings:
- Drop or bridge operator connections upon deletion
- Drop, bridge or keep connections upon disabling
- SSL certificates stored in .RapidMiner/cacert are now trusted on startup. See trust-certificates for more information.
- Added support to open operator tutorial processes directly from the web.
- The "Import Data" dialog for CSV files will try to guess the best matching date format and preselect date for attributes that contain mostly matching date entries
- The "Import Data" dialog for Excel files does now differentiate between date, time and datetime columns specified in Excel
- Improved CSV import wizard to use the structure found in the header or starting row
- Parse Numbers and the Data Import wizards now support exponents in numbers with a leading '+' for positive exponents, e.g. "5.9876E+7"
- Improved Cross Validation error handling when the Performance port is not connected
- The XML Panel does no longer hide default values
- Split thread settings in foreground and background threads (for the currently opened process and processes running in the background, respectively)
- Updated bundled Java for Windows and OS X to version 8u181. This should fix right-click issues on OS X
- Added support for aggregation functions for Pivot operator and improved performance
- When moving operators in the Process view, connected operators will be rearranged and moved to the right if necessary
- For large ExampleSets with more than ~71.5 million rows, the result table will compress the height of each row a bit to accomodate. Data sets with more than ~86 million rows will only display the first ~86 million rows and show a warning that the rest is cut off.
- Fixed an issue that could cause Studio to be stuck for up to ~2 minutes on start-up.
- Fixed very rare process error when working with attribute weights.
- X-Means item count of cluster model will now show the correct size.
- Fixed an issue where (temporary) Access files could not be deleted in a RapidMiner process.
- Added registerLanguage method to the I18N class, which allows to add new languages to the Settings->Preferences->Language selection. The i18n is picked up by providing resource bundles in the usual form of for example GUI_ja.properties and Error_ja.properties. If you want to get a list of not-yet-translated keys, add a file called translation_help.txt in your .RapidMiner folder. After you shut down Studio with your new language selected, it will write all keys for which it did not find the translation in it. This should help you identify keys that you still need to translate.
- Added the OperatorPortActionRegistry to add actions to operator ports.
- Added identifier for last delivering port to the IOObject's userdata via IOObject.getUserData(DeliveringPortManager.LAST_DELIVERING_PORT)
- Added support for parameter dependencies and hidden state to the settings dialog.