Skip to main content
Version: Next

New Features and Improvements

AtScale contains the following new features and improvements.

Support for inbound query CROSSJOIN operations that return empty cells​

AtScale can now be configured to calculate crossjoins of dimensions with empty values. This functionality is controlled by the following engine settings:

  • query.emptyValueCrossJoin.enabled: When enabled (true), AtScale can calculate crossjoins of dimensions that contain empty values. This setting is disabled by default, and does not require an engine restart.

    Important

    Enabling this setting allows the generation of FULL CROSSJOIN queries that WILL be slower and more expensive to run than the corresponding standard join queries that are generated when this setting is disabled.

  • query.emptyValueCrossJoin.requireConstraints: When enabled (true), crossjoins of dimensions that contain empty values (controlled by query.emptyValueCrossJoin.enabled) are only calculated if the query contains constraints. This setting is enabled by default, and does not require an engine restart.

    Important

    Disabling this setting when query.emptyValueCrossJoin.enabled is enabled is not recommended, as it may result in unconstrained crossjoins.

For more information, see Query Settings.

ATSCALE-26138, ATSCALE-26705, ATSCALE-26383, ATSCALE-26264, ATSCALE-27397

New query scopes for semi-additive measures​

You can now configure AtScale to scope queries on semi-additive measures by measure, cube, or dataset. This behavior is controlled by the new query.planning.semiAdditive.multiMeasureHandling engine and cube settings.

Note

The new setting query.planning.semiAdditive.multiMeasureHandling replaces the boolean setting query.planning.semiAdditive.forceIndividualMeasureCheck. Existing configurations are automatically migrated as follows:

  • query.planning.semiAdditive.forceIndividualMeasureCheck = True is migrated to query.planning.semiAdditive.multiMeasureHandling = 'measure'.
  • query.planning.semiAdditive.forceIndividualMeasureCheck = False is migrated to query.planning.semiAdditive.multiMeasureHandling = 'model'.

For more information, see Query Settings (engine-level) and Query Settings (cube-level).

ATSCALE-26895

New query role for data warehouse connections​

AtScale contains a new query role for data warehouse connections, called Aggregate Build Queries. Connections assigned this role handle queries made when building aggregate tables.

Previously, the System Queries query role handled both system queries (e.g. "Stats queries") and aggregate-build queries. Now, the aggregate workload is assignable to a separate data warehouse connection using the new role, effectively allowing stats queries and aggregate queries to run on different connections. This enhancement allows administrators to fine-tune their workload distribution by having stats queries run on small data warehouses and aggregate queries run on larger data warehouses tuned for batch workloads that require longer query time-out settings.

For more information on query roles, see Adding Data Warehouses.

ATSCALE-26125

Updated Snowflake JDBC driver​

AtScale's internal Snowflake JDBC driver has been updated to 3.24.2.

ATSCALE-24922