Azure Stream Analytics JSON Parsing

Question

Trying to stand up a relatively simple ASA query to pull data being ingested as an Array. The piece I cant seem to figure out is how to have the returned table auto populate columns for each Key. Meaning we need each row to have a column for Temperature Value, Humidity Value, Illumination Value, and Contact Value if it exists. Some other arrays will only have Key: Temperature and Key: Humidity.

Thanks!

"telemetry": {
      "data": [
        {
          "key": "temperature",
          "value": 19.1,
          "unit": "°C"
        },
        {
          "key": "humidity",
          "value": 38.5,
          "unit": "%"
        },
        {
          "key": "illumination",
          "value": 0,
          "unit": "lx"
        },
        {
          "key": "accelerationStatus",
          "value": "heartbeat",
          "meaning": "Heartbeat"
        },
        {
          "key": "accelerationX",
          "value": -0.02,
          "unit": "g"
        },
        {
          "key": "accelerationY",
          "value": 0,
          "unit": "g"
        },
        {
          "key": "accelerationZ",
          "value": 1.01,
          "unit": "g"
        },
        {
          "key": "contact",
          "value": "open",
          "meaning": "Window opened"
        }
      ],

Answer 1

You could use UDF in ASA.

UDF function code:

function main(arg) {
    var array = arg.telemetry.data;
    var map = {};
    for(var i=0;i<array.length;i++){        
        var key=array[i].key;
        if(key == "temperature" || key == "humidity" || key == "illumination"){
            map[key] = array[i].value + "" + array[i].unit;
        }
        if(key == "contact"){
             map[key] = array[i].value;
        }    
    }
    return map;
}

SQL:

WITH 
c AS
(
    SELECT 
    udf.processArray(t) AS result
    FROM input AS t
)

SELECT c.result
FROM c

Result: