New analytics and windowing capabilities in Stream Analytics
Published date: October 02, 2020
The latest update of Azure Stream Analytics, deployed in September, includes the following improvements:
SQL Query Language improvements:
- Snapshot Windows: a window type allowing to group by events arriving at the exact same time using System.Timestamp() as grouping key.
- Additionally, System.Timestamp() can now be combined with other windows and will be interpreted as a key in this context.
- TopOne and CollectTop are now available as analytic functions, in addition to the previous aggregate behavior. This will allow to return records ordered by the column of your choice, without the needed to be part of a window. This extends the support for aggregate as stateful analytic functions that was already available with AVG, COUNT, MAX, MIN and SUM.
- Improvement to MATCH_RECOGNIZE : based on customer feedback, we improved our pattern matching features and documentation, allowing customers to build rich CEP (Complex Event Processing) pipelines in a much simpler way.
- New compiler warnings to help to detect some common query issues.
Other product improvements
This update includes hundreds of performance and functional improvements. While we are not releasing full release notes, we wanted to call out the following ones:
- Support of 1MB message to Service Bus output.
- Add Topic as optional property when using EdgeHub as input or output.
- Fixed an issue that prevented to detect large blob in input in some rare cases.