Release Notes - 2020.10.21

Welcome to the 2020.10.21 product release. This article provides an overview of the product innovations and improvements to be delivered on 21 October 2020.

OneAI Forecasting within Storyboards

We are super excited to release a compelling new capability allowing you to run a OneAI Forecast in real-time from within a Storyboard. Check out our user guide for more information on how to use this today in your Storyboards:


Improvements to Data Pipeline Processing

  • Improved the stability of Model Processing by cleaning up the individual tables used when processing the data, instead of cleaning them all up by deleting the schema that was used. This approach is more stable at the Database layer as it means the processing Database Schema is never dropped, and always available. (ref 5832)

Improvements to Data Ingestion

  • Fixed an issue where Redshift Data Destinations restricted the Schema name incorrectly. This would occur when the user set a Data Destination to an External (outside One Model) Redshift Data Destination, and tried to use the same schema name as is used in a One Model Data Source. This restriction now only applies to an Internal Redshift Data Destination, as this loads the data into the same Redshift Database as the Data Sources. (ref 6330)

  • Fixed an issue where in some cases, instead of being able to see and edit a Processing Script for a Data Destination, an ID would appear. This occurred when going to edit a Processing Script that had just been saved, and had to be worked around by refreshing the Data Destinations page.. (ref 6329)

Performance & Platform Improvements

  • We improved the accuracy of metrics that have a multiplication in the Metric Definition when creating/editing metrics. This makes use of a new feature from Amazon in Redshift that allows it to support multiplication of two numbers with a high number of decimal places. (ref 5657)

  • Time Models now consistently return the correct number of time periods when used for non-calendar time dimensions such as Pay Period or Fiscal. Prior to this change, Time Models used date calculations based on a standard Gregorian Calendar, which could lead them to returning an incorrect number of time periods for Time Dimensions that used any non-Gregorian calendar. (ref 6143)


Was this article helpful?

0 out of 0 found this helpful



Please sign in to leave a comment.