Release Notes - 2020.04.01

New User Experience (UX) & Storyboards is Generally Available

Improvement to how Time is Displayed in Data Visualizations

Data Ingestion Bugs Fixed and Minor Improvements

Improvements to Data Pipeline Processing

General Bugs Fixed and Minor Improvements

OneAi Bugs Fixed

It has been a huge project, but we are now at the point where we can say that the new User Experience for One Model is available for general use.  While still finding some small issues and adding functionality, the customer feedback and real-world use has meant we can now move beyond the beta period into mainstream product availability. 

In this release users may experience a small delay on initial login to the new UI.  In moving from beta to production we have forced a reload of the browser cache to ensure the latest content is available for users.  This will be removed over the next couple of releases.

We will keep the traditional user experience and dashboards available for the next 6 months to allow for a smooth transition and give plenty of time for change management for your users and key stakeholders.

Thanks again to everyone for your active enthusiastic collaboration.  Some exciting new capabilities are starting to flow fairly quickly.  Look for the next release where you will be able to add Time as a filter on a Storyboard - a long requested feature for the traditional Dashboards that will become a reality in the new user experience.

We have changed the options in Explore so that you can now use time as a filter only and not have to have it displayed in the data visualization.  Until now Time was required to be displayed in a query, but you can now remove the time display from the Explore query UI and just use time as a filter.  (ref’s include 4223, 4227, 4228, 4229, 4225)

This new capability will add a lot of flexibility to how you can use One Model data and has been requested by a number of customers.  The following screenshots illustrate one example of how to use this new capability can be leveraged.

In the first example you can see a traditional result from filtering on two segments of time, in this case 2016 and 2017 for each Org Unit:

In the example below you see that we can now deselect the Year option from being displayed and the results in the chart have now been filtered for 2016 and 2017.  Since the metric being used in this example is Average Headcount we see the average result across those two years displayed as a single value per Org Unit:

We previously released a feature to allow you to download the result files from One Ai.  In this release we have expanded this capability to include all files when an augmentation has multiple results, i.e. Classification, Regression, Forecast, Commute Time. (ref 4266)

We spent time during the last release cycle to make performance improvements to One AI. One AI now runs in %30 less time.

Fixed an issue where the Oracle connector would fail if Oracle (incorrectly) returned some unicode characters. (ref 4036)

Fixed an issue where incorrect credentials in the SuccessFactors Connector could lock the API run into "Waiting to Start". (ref 4039)

When a new user is created (and assigned a Data Access Role), or assigned a new Data Access Role, any Storyboards associated with that role will be automatically cached for the user. These queries will be cached immediately, and automatically re-cached for 6 weeks following that point.  The 6 week period resets each time a user opens a storyboard. This new caching capability, is a feature of the new Storyboards capability as part of the new User Experience.  

To enable Automated Caching, enable the Use Automated Cache Warming flag in the Company Page.  This new capability complements the existing features to identify specific Dashboards/Storyboards to be cached, but adds additional automation logic based on user activity.

When updating on-site Metrics and Dimensions at the end of Processing scripts, we will now automatically cancel any queries that are running at that time so they don't delay the update. This may cause some queries running at that point in time to error unexpectedly, but these queries can then be re-run straight away. This will also affect Direct Connection queries to Redshift.  (ref 3506)

Fixed a bug that caused the EDA report to not render when a label has extremely low cardinality.

 Fixed a bug related to composite dataset ID that caused runs to fail while using the new key with three unique dataframes.