Azure流分析查询中的限制1或前1

Question

我在下面写了Stream Analytics Query以获取设备离线状态警报。如果在最后5分钟内没有从设备收到任何输入，查询会给我一个警报。

我需要该设备的最后一个远程信息处理记录（最后已知的位置），但我正在获取为该设备发送的所有数据。我尝试了LAST和TopOne，但都是开窗方法。除了自定义功能之外还有其他解决方案吗？

SELECT
     t1.header.serialNumber as serialNumber,t1.header.make as 
     make,t1.header.messageTimestamp as MessageTime,'Device Offline Alert' as 
     alertType
INTO
[alertOutput2]
FROM
  [tsfInput] t1 TIMESTAMP BY header.messageTimestamp
LEFT OUTER JOIN [tsfInput] t2 TIMESTAMP BY header.messageTimestamp
ON
     t1.header.serialNumber=t2.header.serialNumber AND 
     t1.header.make=t2.header.make
     AND DATEDIFF(minute, t1, t2) BETWEEN 1 and 5
     WHERE t2.serialNumber IS NULL

样本输入

{
"header": {
    "assetKey": 8910053376183227000,
    "make": "CAT36",
    "serialNumber": "Truck36",
    "deviceType": "A5:N2",
    "isPriority": false,
    "isReplay": true
}
}

Answer 1

我能够修复查询但不是100％。从查询中删除以下行并且它正在工作之后，仍然不清楚它为什么不使用TimeStamp

TIMESTAMP BY header.message时间戳

我有2个查询1设备在线警报和第2个设备离线警报

SELECT
*,'Device Online Alert' as alertType
INTO
    [alertOutput]
FROM
[tsfInput] 
WHERE ISFIRST(mi, 5) OVER (PARTITION BY header.serialNumber,header.make) = 1

SELECT t1.*,'Device Offline Alert' as alertType 
INTO 
 [alertOutput2] 
FROM [tsfInput] t1 
 LEFT OUTER JOIN [tsfInput] t2 
ON t1.header.serialNumber=t2.header.serialNumber AND t1.header.make=t2.header.make 
AND DATEDIFF(minute, t1, t2) BETWEEN 1 and 5 
WHERE t2.header IS NULL