Oracle Data Miner 4.0.2 (Oracle SQL Developer 4.0.2):  Bug Fixes  

Oracle Data Miner 4.0.2 adds a number of minor fixes. Below are the fixes relevant to users. These features are supported for all database versions supported by Data Miner (Oracle Database 11.2.0.1 and above).

  • SQL Developer 4.0.2 corrected an issue on how Oracle NLS client side libraries were managed.  The problem in 4.0.1 resulted in Data Miner repository installation failures on databases requiring national language support (NLS) client side libraries to be loaded.    This occurred only when the NLS numerical format settings used a comma in place of a decimal.  SQL Developer 4.0.2 fixes this problem with Oracle Data Miner installations on databases that required local national language support (NLS) libraries. 

  • Oracle Data Miner 4.0.2 prevents user from attempting to install Data Miner repository in CD$ROOT container database (Oracle Database 12c issue only).
  • Model Build Nodes were failing when a CLOB was passed in as input.  This only occurred on 11.2 databases.

  • Aggregation Nodes:
    • Date/TimeStamp columns are now removed from columns that can be selected for aggregation because aggregation of Date/TimeStamp data is not supported.
    • The Aggregation node included a NOT NULL clause to reduce the number of null DM_NESTED types generated for the sub group option.  This restriction has been removed.
  • GLM models now allow both Ridge Regression and Feature Selection to be disabled.  Additionally there is a fix to the Model Viewer so it properly displays generated feature expressions. The node algorithms editor also has an improved layout for the algorithm settings editor.
  • Transformation Nodes:
    • When Opening a Transformation node that has been run, the first row of the table was not selected, resulting in  the histogram not being loaded in the panel below.  
    • Editing a custom transformation required the user to change the name of the generated column but this requirement has been removed.
    • Corrected an issue that would result in Node failure when generating Histograms.  The failure was triggered by longer character values. This also occurred in the Explore Node.
  • The Join Node data types were not being properly handled.  This was preventing some join specifications from being accepted.
  • Apply Node fixes a problem where after changing the model target column data type, the connected Apply node did not properly refresh generated predictive output columns.

JDK Version Validation Applied to Data Miner 4.0.2

  • There is a minimal version of JDK that is required by SQL Dev. A warning or error is displayed depending on the severity of the version difference. There is, unfortunately, an error in the warning message, that refers the user to an incorrect control file to adjust the JDK that SQL Dev is referencing. Please see the SQL Dev documentation on JDK management to understand the correct procedures and also this posting 4.0.2 doesn't start : Unsupported java version... on the SQL Developer Discussion Forum thread for more information.

For More Information and Support, please visit the Oracle Data Mining Discussion Forum on the Oracle Technology Network (OTN)

Return to Oracle Data Miner page on OTN



In-Memory Replay Banner