November 22, 2022: What’s New in Qlik Cloud?

Direct Query now supports Google BigQuery

You can now connect to Google BigQuery databases using Direct Query. Direct Query allows analytics applications to directly query cloud databases with SQL pushdown as users interact with data through visualizations and filtering. App developers can use Direct Query to build SQL-centric applications for big-data analysis or near real-time scenarios.

Accessing cloud databases directly with Direct Query

New feature insights in AutoML

AutoML has new insights to help clarify what happens when the machine learning algorithms train on a dataset: the underrepresented class insight and the one-hot encoding insight. Underrepresented class insight is when a value in the column has less than ten instances (only in the binary case). A one-hot encoding insight is when a categorical feature has less than 14 unique values. If more than 100 columns are selected for an experiment then all the categorical features are impact encoded, regardless of how many unique values they have.

Categorical encoding

Configuring experiments

Updates to font styling for tables and pivot table

App developers can now style the font used for titles, subtitles, and footnotes in tables and pivot tables, giving them more options for customizing their visualizations. A new General tab provides options to change the font type, size, color, and emphasis. The Chart tab controls styling of rows, scrollbars, and custom headers.

Table

Pivot table

Process type 2 dimensions and preserve change history in data marts

Data Mart tasks can now incrementally process type 2 dimensions and preserve the change history of a slowly changing dimension. It works by storing and managing current and historical data values in the data warehouse over time. Incremental processing provides fully automated ELT instructions which are able to detect changes across multiple source tables as part of denormalization to reduce processing time, cost and improve accuracy. Additionally late arriving dimensions are also automatically handled by the data mart task.

Creating and managing data marts