在 Azure 流分析查询中重用子查询的结果

Posted 2023-05-08

【中文标题】在 Azure 流分析查询中重用子查询的结果

【英文标题】：Reuse the result from subquery in Azure Stream Analytics Query

【发布时间】：2017-11-29 15:41:08

【问题描述】：

我正在使用 Azure 流分析，但在查询部分遇到了一些问题。这是我的代码。

WITH subquery as (
    SELECT 
    messageId,
    deviceId,
    temperature, 
    humidity,
    EventProcessedUtcTime,
    DemoML(temperature, humidity) as result1
    from DemoInput
    )

SELECT
    messageId as messageId,
    deviceId as deviceId,
    temperature as temperature,
    humidity as humidity,
    EventProcessedUtcTime as EventProcessedUtcTime,
    result1.[Scored Labels] as result,
    result1.[Scored Probabilities] as resultProbability
INTO
    [DemoOutput]
FROM
    [subquery]


SELECT
    result1
INTO
    [c2d]
FROM
    [subquery] 

DemoML 是一个返回结果的函数。我想将 result1 放入两个不同的输出中。但我只设法将 result1 放入一个输出中。我怎样才能做到这一点？我对 SQL 完全陌生。

【参考方案1】：

您可以使用INSERT ... OUTPUT ... INTO ... SELECT 插入两次：

WITH subquery as (
    SELECT 
    messageId,
    deviceId,
    temperature, 
    humidity,
    EventProcessedUtcTime,
    DemoML(temperature, humidity) as result1
    from DemoInput
    )
INSERT INTO target_1(col1,...)
OUTPUT inserted.col1, ...
INTO target_2(col1, ..)
SELECT
    messageId as messageId,
    deviceId as deviceId,
    temperature as temperature,
    humidity as humidity,
    EventProcessedUtcTime as EventProcessedUtcTime,
    result1.[Scored Labels] as result,
    result1.[Scored Probabilities] as resultProbability
FROM  [subquery];

简化：

CREATE TABLE t1(i INT);
CREATE TABLE t2(i INT);
CREATE TABLE src(i INT);
INSERT INTO src(i) VALUES(10),(20);

INSERT INTO t1(i)                 -- target one
OUTPUT inserted.i
INTO t2(i)                        -- target two
SELECT i
FROM src;

DBFiddle Demo

【讨论】：

在这种情况下，我是否将相同的数据放入两个不同的目标？

@SamTew 是的，只需检查演示链接，但如果需要，您可以操作列列表。

如果我想把两组不同的数据放到两个表中怎么办？

@SamTew 然后你可以使用INSERT over DML 语法。你原来的问题已经解决了。

@SamTew 我在这里看到的唯一区别是 Azure ML 调用。您是否对它进行了独立测试以确保它返回一些值？