Azure Stream 具有多个输入的分析查询

Question

Issue

I have a Azure IoT-Hub sending messages to a Azure Stream Analytics Job.

Each message contains a 'NodeName'. I have a table 'plcnext_nodes' which has a unique 'NodeId' for each node with their corresponding 'NodeName'.

How can I use both the input from the 'plcnext_nodes' table and the IoT-Hub messages to store the event data in another SQL table using the 'NodeId'?

I want to use the 'NodeId' instead of the 'NodeName' because some names can get very long and saving them over and over with each message is a waste of storage.

---

Desired solution

I would like to parse the following message from the IoT Hub :

{
    "NodeName": "ns=5;s=Arp.Plc.Eclr/DI2",
    "NodeDataType": "Boolean",
    "EventValue": 0,
    "EventMeasuredUtcTime": "2019-11-11T12:15:22.4830000Z",
    "EventProcessedUtcTime": "2019-11-11T12:41:57.1706596Z",
    "EventEnqueuedUtcTime": "2019-11-11T12:15:32.1260000Z",
    "IoTHub": {
        ...
    }
}

Compare the 'NodeName' with those in the plcnext_nodes table to get the appropriate 'NodeId':

NodeId  NodeName                 NodeDataType
---------------------------------------------
1       ns=5;s=Arp.Plc.Eclr/DI1  Boolean
2       ns=5;s=Arp.Plc.Eclr/DI2  Boolean
...

To get the following output and insert into the plcnext_events table:

NodeId  EventValue  EventMeasured
-----------------------------------------------
1       0           2019-11-11 12:15:22.4830000

---

Query

I have tried the following query on Azure Stream Analytics:

SELECT
    NodeId,
    EventValue,
    EventMeasuredUtcTime,
    EventEnqueuedUtcTime,
    EventProcessedUtcTime
INTO
    [plcnext_events]
FROM
    [plcnext_nodes],
    [iot_hub]
WHERE
    [iot_hub].NodeName = [plcnext-nodes].NodeName

But JOIN is not supported in FROM and I haven't been able to use a JOIN clause because of it's DATEDIFF restriction (The plcnext_nodes table has no timestamps)

Is there a way to achieve this?

Answer 1

You can use a Reference Data JOIN in Stream Analytics to join tables.

Using Reference Data for Lookups in Stream Analytics

Reference data (also known as a lookup table) is a finite data set that is static or slowly changing in nature, used to perform a lookup or to augment your data streams.

In your case, go under the Inputs section in your Stream Analytics job and add a Reference Input. You can select either Blob Storage or SQL Database as your source.

You can then define your SQL query to return the reference data you need. For your case, your reference data query would look like this:

SELECT NodeId, NodeName, NodeDataType
FROM dbo.plcnext-nodes

Once you define your reference data join in SA, go to the Query section in Stream Analytics and update your query. For your scenario your query would look like this (with the JOIN):

SELECT
   pn.NodeId,
   hub.EventValue,
   hub.EventMeasuredUtcTime,
   hub.EventEnqueuedUtcTime,
   hub.EventProcessedUtcTime
   INTO [plcnext_events]
FROM [iot_hub] hub
JOIN [plcnext-nodes] pn ON pn.NodeName = hub.NodeName