SSAS 时间序列,预测何时会出现故障
我想使用 SSAS 和 TimeSeries 挖掘结构来预测可预测值何时达到某个阈值。
例如:
SELECT [Info Key],
PredictTimeSeries([Free Space], 200) as ForcastedSize
FROM [Drive Module Information]
WHERE ForcastedSize < 10000 --(<< this does not work)
这将告诉我预测驱动器空间将低于 10000 的日期。
如何编写 MDX 查询来完成此操作? 谢谢, 布莱恩
更新1: 我想我可以通过这种方式完成它,但有一些限制:
SELECT [Drive Module Information].[Info Key],
(SELECT *
FROM PredictTimeSeries([Drive Module Information].[Free Space], 5000) as [FUTURE]
) AS T
FROM [Drive Module Information]
WHERE
[Info Key] = 'MyMachine C:' AND
[Free Space] < 10000
限制是我只能向前看 X# 步而不发疯。这没关系。我很高兴知道驱动器在下周或下个月不会填满。
我不知道在这种情况下如何使用 FILTER,并且仍然好奇是否存在“此可预测值将在什么日期等于此值”。
更新2:我得出的结论是SSAS不应该这样做,所以在我找到不同的答案之前,我会将icCube标记为答案,因为他提供了帮助。
I would like to use SSAS and a TimeSeries Mining Structure to predict when the predictable value will reach a certain threshold.
For Example:
SELECT [Info Key],
PredictTimeSeries([Free Space], 200) as ForcastedSize
FROM [Drive Module Information]
WHERE ForcastedSize < 10000 --(<< this does not work)
This will tell me the date that it forecasts that the drive space will be below 10000.
How do I write an MDX query to accomplish this?
Thanks,
Brian
UPDATE 1:
I think I can accomplish it this way, with some limitations:
SELECT [Drive Module Information].[Info Key],
(SELECT *
FROM PredictTimeSeries([Drive Module Information].[Free Space], 5000) as [FUTURE]
) AS T
FROM [Drive Module Information]
WHERE
[Info Key] = 'MyMachine C:' AND
[Free Space] < 10000
The limitation is that I can only look X# of steps forward without getting crazy. Which is ok. I am ok with knowing that the drive will not fill up over the next week, or month.
I did not figure out how to use FILTER in this situation, and am still curious as to if there is a "What date will this predictable value be equal to this value".
UPDATE 2: I have come to the conclusion that SSAS was not meant to do this, so until I find out differently, I will mark icCube as the answer since he helped out.
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。

绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(1)
MDX 不是 SQL,MDX 的 where 子句不是真正的过滤器。作为快速介绍,您可以浏览一下MDX 温和教程。
您可以使用MDX 过滤器功能。
MDX is not SQL, the where clause of MDX is not a real filter. As a quick introduction you can go through this MDX gentle Tutorial.
There is a MDX Filter function you can use.