Superset 2.0 is a big step forward. This release cleans up many legacy code paths and feature flags, and deprecates lots of legacy behaviors in Superset.
-
Addition of a statsd gauge metric for Slack and email notifications for increased visibility into errors around alerts / reports (#20158)
-
Helm chart now supports resource limits and requests for each component (#20052)
-
New GitHub workflow to test Storybook Netlify instance nightly (#19852)
-
Minimum requirement for Superset is now Python 3.8 (#19017
Charting and Dashboard Experience
Support for horizontal bar chart added (#19918)
Time Series Charts now support stacking of both negative and positive values (#20408)
-
Pie charts now defaults to a row limit of 100 to prevent crashes when a high-cardinality column is chosen as the dimension (#20392)
-
World map chart now supports coloring either by metric or by the country column (#19881)
-
Table visualization now supports drag and drop for columns (#19381)
-
Mixed chart now supports Advanced Analytics (#19851)
-
Add support for generic x-axis (non-time-series) in the Mixed Chart (#20097)
- Charts can now be created in Edit Dashboard mode (#20126)
-
Add aggregate total for Pie charts (#19622)
-
Legend is now enabled by default for relevant charts (#19927)
View Results Experience
-
Explore and Dashboard views now support displaying of multiple results (#20277)
-
Results pane in Dashboard view now more closely mirrors rich functionality from Results pane in Explore (#20144)
Quality of Life
-
Edit Dataset modal now doesn't close when you click away (#20278)
-
When editing the label in the Metrics popover, pressing Enter now saves what you typed (#19898)
-
When adding new chart from the dashboard view, the dashboard name will now pre-fill (#20129)
-
Annotations now have an improved empty state (#20160)
-
Confirmational modal is now shown if user exits Explore without saving changes (#19993)
-
Explore popovers now close when the Escape key is pressed (#19902)
-
Run and Save buttons are redesigned for improved usability (#19558)
Databases
-
Native database driver for Databricks (#20320)
-
Time grains for SQLite are now simplified (#19745)
-
Multiple upgrades to the Trino database engine (#20152)
-
Switch from sqlalchemy-trino to trino-python-client (#19957)
-
Apache Pinot now supports more time grains in Superset (#19724)
Jinja
-
New Jinja macro enables querying / referencing both physical and virtual datasets in SQL Lab (#15241)
-
New Jinja macro added to improve experience of including multiple items (#19574)
Other
- Datasets can now be filtered by their certification status (#20136)
-
Initial implementation of advanced types (#18794)
- Flag:
ENABLE_ADVANCED_DATA_TYPES
- Flag:
-
Caching can now be enabled in database setups when user impersonation is enabled (#20114)
- Flag:
CACHE_IMPERSONATION
- Flag:
-
Control behavior for how color palettes are chosen (#19987)
- Flag:
USE_ANALAGOUS_COLORS
- Flag:
-
Enabling non-time-series x-axis in some charts (#20097)
- Flag:
GENERIC_CHART_AXES
- Flag:
-
As part of the 2.0 cleanup efforts, the following feature flags were removed (which means the behavior is now permanently enabled):
ROW_LEVEL_SECURITY
ENABLE_REACT_CRUD_VIEWS
PUBLIC_ROLE_LIKE_GAMMA
-
The following feature flags are now True by default, but can be turned back to False:
ENABLE_EXPLORE_DRAG_AND_DROP
ENABLE_DND_WITH_CLICK_UX
DISABLE_LEGACY_DATASOURCE_EDITOR
SQLLAB_BACKEND_PERSISTENCE
VERSIONED_EXPORT
-
The following config flags were removed:
APP_ICON_WIDTH
-
A number of legacy interim flags were removed around SIP-15 (#18936)
-
The
ENABLE_JAVASCRIPT_CONTROLS
flag was moved from an app config to a feature flag
To learn more about the breaking changes in 2.0, please read UPDATING.MD