Hi, Azure Stream Analytics is optimized for streams of data coming in real time. However we can use some workaround for processing "historical data" coming to blobs.
By default, Azure Stream Analytics uses the arrival time as timestamp. For blob, you are right this is the LastModified date. However using the keyword TIMESTAMP BY you can choose another timestamp from your payload.
Note that you will need to extend the late arrival policy in the options so the difference between the actual timestamp and the arrival time is less than the maximum "late arrival time".
Late arrival can be disabled totally to support reprocessing of historical data by setting the option to "-1", however this is not possible to do this in the portal and you will need to do this either using the API, and Visual Studio (e.g. export the job, edit the number, and republish the job).
Let us know if you have further question or need additional help.
JS (Azure Stream Analytics)
PS: See more info on time policies here: https://docs.microsoft.com/en-us/azure/stream-analytics/stream-analytics-time-handling
Hi, thanks for reporting this. Sorry for the late response.
Other types, like DATETIME do work, but only if there is an explicit CAST or CREATE TABLE.
i.e. you would have to do “min(cast (d as datetime))” vs. just “min(d)”
We'll update the doc with this information.
Hi, this is a supported scenario. Do you have any issue when running your job? Can you share more info here?
22 votesstarted · 2 comments · Stream Analytics » Inputs and Outputs · Flag idea as inappropriate… · Admin →
REPLACE is now available in the product: https://docs.microsoft.com/en-us/stream-analytics-query/replace-azure-stream-analytics
(Editing the title to reflect this change and track other functions)
Thank you for the feedback Krishna. We are constantly extending set of built-in functions and will consider these functions in one of the next iterations. This is the list of currently supported string functions: https://msdn.microsoft.com/en-us/library/azure/dn931798.aspx
28 votesunder review · 2 comments · Stream Analytics » Query Language · Flag idea as inappropriate… · Admin →
The ROUND method had been added to the product. The documentation is here: https://docs.microsoft.com/en-us/stream-analytics-query/round-azure-stream-analytics
I'll rename this suggestion to track the "split" function that is not yet available.
114 votesunder review · AdminAzure Stream Analytics Team on UserVoice (Product Manager, Microsoft Azure) responded
we will consider this request.
Thanks for the feedbacks. In order to reduce cost, it is possible to use 1 SU for small jobs (this is about $80 per month).
1 SU jobs are able to have different sources, sinks and query steps. So it's actually possible to pack various queries inside 1 job.
By user timestamped reference data, the reference data will be refreshed when the job is running.
We announced MSI auth for ADLS: https://docs.microsoft.com/en-us/azure/stream-analytics/stream-analytics-managed-identities-adls
More to come...
A workaround for this would be to create 2 jobs:
- A first job for the critical output (keeping the error policy to "retry")
- A second job for non-critical output (changing the policy to "drop")
60 votesunder review · AdminAzure Stream Analytics Team on UserVoice (Product Manager, Microsoft Azure) responded
Reopening this item since we have additional votes and heard about more comments related to this.
Few comments/workaround in the meantime:
- For PII/HBI data: ASA on IoT Edge and Azure Stack for PII – this allows pre-processing at the Edge without sending data to the cloud.
- Streaming + DB can work side-by-side on ASDE (Azure SQL Database Edge)
Thank you for your interest in using Stream Analytics. We are looking for more details on this scenario. For those of you that are interested in an input for On-prem SQL, are you looking to use the SQL data as reference data to be joined with another incoming data stream? Or looking to using the SQL data as a data stream? In the reference data case, you can use Azure Data Factory to take periodic snapshots of the relevant reference data and copy it to blobs that can then be used by Azure Stream Analytics.
For those of you looking for a direct output to On-Prem SQL, while we don't yet support this scenario directly, it is possible as a workaround to use Azure SQL DB as an output from Stream Analytics and setup a scheduled data move using Azure Data Factory to move the data from Azure DB to your On-Prem SQL database and accomplish a similar pattern.
Thank you for reaching out Hamid.
Can you clarify? Are you suggesting that Azure Stream Analytics directly pull from sources such as the Twitter firehose? What would pulling from a device mean?
Does DateTimeFromParts work for you?
If not, then please let us know why not.
1 voteunder review · AdminAzure Stream Analytics Team on UserVoice (Product Manager, Microsoft Azure) responded
Relay has its own characteristics and advantages, but you may be able to use Servicebus Topics or Queues instead, which are supported by ASA today.
Thank you Gökhan for the reply and valuable feedback. Let us look into this suggestion.
While we are considering further output targets, please note that we support Azure Storage Tables and Azure DocumentDB today - both of which qualify as NoSQL stores.