Azure Stream Analytics is a fully managed service allowing you to gain insights and run analytics in near real-time on your big data streaming workloads. The service was first deployed more than 2 years ago, long before the “new” Azure management portal, https://portal.azure.com, even existed.
For the past few months we’ve been hard at work adding exciting new features to the service as well as transitioning the management user interface from the old https://manage.windowsazure.com to the new portal
Today we want to announce that we’ve just added the ability to test queries in the “new” portal without needing to start or stop a job. Here’s a quick look at how this works.
You can setup a Stream Analytics by following this simple tutorial - How to create a Stream Analytics job.
Once you have created a new Stream Analytics job you would typically Create Inputs and then Create Outputs. Or you can just skip ahead to building the query and once your query is working then go back and define the Inputs and Outputs to match those used in the query. Both ways work, giving you the flexibility to decide how you wish to work.
For the purposes of this blog post I have defined a job with 1 data stream input, called StreamInput and 1 output, called Output. You can see these in the query editor blade above.
Open the Query editor blade from the job details screen by clicking on the query in the “Query” lens. Or in our case the < > placeholder because there is no query yet.
You will be presented with the rich editor as before where you create your query. This blade has now been enhanced with a new pane on the left. This new pane shows the Inputs and Outputs used by the Query, and those defined for this job.
There is also 1 additional Input and Output shown which I did not define. These come from the new query template that we start off with. These will change, or even disappear all together, as we edit the query. You can safely ignore them for now.
A key requirement and a common ask from our customers while writing a query is being able to test, and test often, to ensure that the output is what it is expected to be, given some input data. Having to save the query after every edit, start the job, wait for incoming data, check the results, and then stop the job again each time you make a small change to the query would be slow and is sometimes not even possible. A way to test changes to a query quickly was needed.
I am happy to announce that with today’s latest release in the portal you can now test the query without going through this stop/start process. Here's how ...
Sample data and testing queries
To test with sample input data, right click on any of your Inputs and choose to Upload sample data from file.
Once the upload has completes you can then use the Test button to test this query against the sample data you have just provided.
The output of your query is displayed in the browser, with a link to Download results should you wish to save the test output for later use. You can now easily and iteratively modify your query, and test repeatedly to see how the output changes.
In the diagram above you can see how I have changed the query inline to have a 2nd output, called HighAvgTempOutput where I am only writing a subset of the data being received.
With multiple outputs used in a query you can see the results for both outputs separately and easily toggle between them.
Once you are happy with the results in the browser, then you can save your query, start your job, sit back and watch the magic of Stream Analytics happen for you.
Feature Parity and the road ahead
With the long awaited addition of sample data and query testing in the new portal we are happy to announce that we have reached feature parity between the portals. Everything you could do before, and more, is now in the new portal. Going forward all new development efforts will be concentrated on the new portal. The old portal will continue to work and existing functionality will remain until end of the calendar year when we place to completely retire support for Stream Analytics in the old portal.
If you have not tried Stream Analytics in the new portal we encourage you to head over and give it a try.
We’re really excited to bring local testing to the new portal and take this final step to reaching feature parity across the two portals. We hope this makes your life much easier as you go about developing (and testing) your queries.
We invite you to provide feedback on our User Voice page about what you want added next to the service!