Grafana Datediff with Synapse Datasource

问：

我有一个简单的查询，可以使用其内置的 MySql Server 在 Azure Synapse 中工作：

SELECT TOP (15)
  cast ("Betriebstag" as DATETIME),
  CONVERT (VARCHAR, "Betriebstag", 34) as "betriebstag",
  "MWNR",
  "kwhUsed" as "Ist-Verbrauch",
  "kWh_prediction" as "ML Prognose",
  "kwh_prediction_linear_temperature" as "Lineare Prognose",
  cast ("prediction_written" as DATETIME)
FROM [dbo].[XYZ]
WHERE ("Betriebstag" BETWEEN '2023-08-01' AND '2023-08-15') 
AND "MWNR" = 4021
AND DATEDIFF(day, "Betriebstag", "prediction_written") = -1 
AND kwhUsed is not null
ORDER BY "Betriebstag" asc

在 Grafana 中，我使用相同的数据库。但是，使用相同的查询时，我收到错误：

convert frame from rows error: mssql: Conversion failed when converting date and/or time from character string.

在 Grafana 中使用 DATEDIFF 有什么区别吗？或者将我的时间戳变量转换为 DATETIME 是否有问题？

谢谢！