Special Attention Notes and Known Issues
The section describes special attention notes that indicate changes in functionality that require you to update models, cause model results to differ, or display a warning message when you first load a model in RiverWare Version 8.5.
If you have any questions, contact RiverWare Support:
Analyze Ensemble Data Set Unit Changes
Within the Analyze Ensemble Data Set script action, a change was made to use internal units in all statistical computations. In some cases, this could lead to slightly different results. See Analyze Ensemble Data Set in Automation Tools for documentation on this action.
Analyze Ensemble Data Set Percentile Change
Previously for the Analyze Ensemble Data Set script action, when Percentile was selected for the Analysis Type, in some cases, it could return a value that was ranked one below the value that should have been returned. This issue was filed as RW-6738 and has been fixed. This fix could lead to numerical differences in model results that use this script action. See Analyze Ensemble Data Set in Automation Tools for documentation on this action.
DSS DMI output changes
Compared to previous versions of RiverWare, in certain situations, different results could be output from DSS DMIs. In RiverWare 8.3, the DSS Database DMI was re-implemented. In the process, a change was made to how RiverWare reads missing values or writes NaN values. In the 8.4.3 patch and 8.5 release, the behavior has been reverted to the pre-8.3 approach.
HDB Database DMI Oracle version
The HDB Database DMI connection was upgraded from Oracle 18c to Oracle 19c. With this upgrade, HDB connectivity requires either Oracle Client 19 or Oracle Instant Client 19. CADSWES tested with Oracle Client 19.3 and Oracle Instant Client 19.15. For more information on HDB datasets, see HDB Database in Data Management Interface (DMI).
Scalar Slots with Expression Evaluation
A scalar slot with expression previously did not clear automatically when evaluated during a run. Thus, in cases that it should have had NaN, it retained a leftover value from earlier evaluations. To fix this, Scalar Slots are now reset/cleared before evaluating. Conceivably, this could lead to numerical differences although none were found in testing.
Revised: 07/05/2022