Apache Superset 1.5: Release Notes
Apache Superset 1.5 is out! Thank you to everyone who contributed to this fantastic release.
Note that Preset Cloud has been running Superset 1.5 plus some add-ons for a few months now!
Faster Loading of Complex Dashboards
Previously, it wasn't uncommon for dashboards with lots of native filters and charts to take 15 seconds to render. Note that these changes only apply to dashboard native filters (not the soon-to-be-deprecated Filter Box component).
Would you like some proof? Check out the before and after videos for the same, complex dashboard!
Before (~11 seconds)
After (~3 seconds)
Cleaner Design for Buttons in Dashboard Filters
Dashboard native filters now sport a cleaner design, with the most notable changes being the relocating of the Apply Filters and Clear All buttons to the bottom of the filter panel.
Dependencies for Dashboard Filters
Dashboard filters just become more aware. Of each other! You can now restrict the exposed values in a filter based on the selected values of other filters in the same dashboard.
More Support for Custom SQL in Explore
In addition to being able to write custom SQL for adhoc Metrics and Filters in Explore, the Column control now also features a Custom SQL tab. This makes it possible to write custom expressions directly in charts without adding them to the dataset as saved expressions.
Improvements to the Caching Experience
Superset 1.5 introduces SupersetMetastoreCache, which makes it possible to cache data in the Superset metadata store without relying on an external cache like Redis or Memcached. You can read more in the Superset caching documentation.
Dashboards with lots of filters used to cause errors because the filter state lived in the URL itself, which resulted in the long URL problem. Now, Superset stores Dashboard and Explore state in the cache (instead of all in the URL).
Building on both of the above features, permanent URL's to Dashboard and Explore pages now store state in JSON objects in the Superset metadata store (instead of the URL). In addition, URL's incorporate hashes instead of numerical id's for added security.
Superset 1.5 contains an incredibly long list of bug fixes and we recommend reading the Fixes section in the changelog to get the full picture. Here's a preview of some of them!
- Fixes occasional blank state issue when only one column is selected (#19651)
- Fixes issue with dashboard crashing when Big Number chart overflows (#19688)
- Fixes issue where a jinja macro wasn't accessible from Metrics in Explore (#19565)
- Remove gap from bottom of SQL Lab results (#19138)
- Fixes to dataset search during chart creation workflow (#19065)
DuckDB is like SQLite, but optimized for OLAP (column-oriented) operations! Here's a link to the PR.
Azure Data Explorer
Azure Data Explorer is a fast, fully managed data analytics service for real-time analysis on large volumes of data streaming from applications, websites, IoT devices, and more.
Here's a link to the PR.
Existing supported databases got nice upgrades as well in 1.5!
- Support for server certificates (PR #16346)
- Support for alternative authentication methods like Kerberos, JWT, and custom auth (PR #17593)
Microsoft SQL Server
- Supports using common table expressions (CTE's) in virtual tables (PR #18567)
- Supports TOP syntax (PR #18746)
- Supports TOP syntax (PR #18240)
- User imperation now possible through another driver (PR #19252)
You have a few different options for upgrading to v1.5:
- Python package: apache-superset
- Source release from Github
- Source release from the [Apache Software Foundation])(https://downloads.apache.org/superset/1.5.0/)
- Docker Image
As usual, you can read the full changelog over in the Superset Github repo.
Notes on Updating
You can find all of the breaking changes, potential downtime changes, deprecations, and other crucial changes in updating.md.