在流分析查询中生成倒置时间戳

Posted 2023-05-08

【中文标题】在流分析查询中生成倒置时间戳

【英文标题】：Generate inverted timestamp in stream analytics query

【发布时间】：2018-07-24 01:46:40

【问题描述】：

我想将流分析作业的输出通过管道传输到表存储。对于行键，我想使用倒置时间戳。在 C# 中，我只需执行以下操作：

var invertedTimestamp = DateTime.MaxValue.Ticks - DateTime.UtcNow.Ticks;

我正在尝试编写 ASA 工作查询，但不知道如何获得类似的内容：

SELECT
    CONCAT(deviceId, '_', sessionId) as DeviceSession,
    DATEDIFF ( ms , NOW(), MAX_DATE() ) //pseudocode
INTO
    tablestorage
FROM
    iothub

【参考方案1】：

您可以使用User Defined Functions

带功能：

/ There are 10000 ticks in a millisecond.
// And there are 621355968000000000 ticks between 1st Jan 0001 and 1st Jan 1970.
function tics_to_max() 
	var convertToTicks = function (date) 
		return ((date.getTime() * 10000) + 621355968000000000);
	;

	var max = new Date('9999-12-31T23:59:59.999Z');
	var maxTics = convertToTicks(max);
	var currentTics = convertToTicks(new Date());

	return maxTics - currentTics;

您的查询可能如下所示：

SELECT
  CONCAT(deviceId, '_', sessionId) as DeviceSession,
  udf.tics_to_max() AS TicsToMax
INTO
  tablestorage
FROM
  iothub

【参考方案2】：

试试下面的代码，它会工作的。

Calendar cal = Calendar.getInstance();
SimpleDateFormat sdf = new SimpleDateFormat("MMM dd yyyy HH:mm:ss", 
Locale.ENGLISH);
try 
    
    String datestart="June 14 2018 16:02:37";
    cal.setTime(sdf.parse(datestart));// all done
     Calendar cal1=Calendar.getInstance();
    String formatted = sdf.format(cal1.getTime());//formatted date as i want
    cal1.setTime(sdf.parse(formatted));// all done
    long msDiff = cal1.getTimeInMillis() - cal.getTimeInMillis();
    long daysDiff = TimeUnit.MILLISECONDS.toDays(msDiff);
    Toast.makeText(this, "days="+daysDiff, Toast.LENGTH_SHORT).show();
     
    catch (ParseException e) 
    
    e.printStackTrace();