在 Snowflake 中使用 CTE 进行 MATCH_RECOGNIZE答案

【问题标题】：MATCH_RECOGNIZE with CTE in Snowflake在 Snowflake 中使用 CTE 进行 MATCH_RECOGNIZE
【发布时间】：2021-07-27 12:10:41
【问题描述】：

我在带有几个 CTE 的查询中使用 MATCH_RECOGNIZE 函数。运行查询时，出现以下错误：

SQL 编译错误：此上下文不支持 MATCH_RECOGNIZE。

在我的查询中，MATCH_RECOGNIZE 前后有几个 CTE，如下所示。

WITH cte1 AS (
SELECT *
FROM dataset
WHERE ID IS NOT NULL AND STATUS IS NOT NULL ),

cte2 AS (
SELECT *
FROM cte1
QUALIFY FIRST_VALUE(STATUS) OVER (PARTITION BY ID ORDER BY CREATED_AT) = 'created' )

mr as (
SELECT *
    FROM cte2
    MATCH_RECOGNIZE (
      PARTITION BY ID
      ORDER BY CREATED_AT
      MEASURES MATCH_NUMBER() AS mn,
               MATCH_SEQUENCE_NUMBER AS msn
      ALL ROWS PER MATCH
      PATTERN (c+m+)
      DEFINE
         c AS status='created'
        ,m AS status='missing_info'
        ,p AS status='pending'
    ) m1
    QUALIFY (ROW_NUMBER() OVER(PARTITION BY mn, ID ORDER BY msn) = 1)
          OR(ROW_NUMBER() OVER(PARTITION BY mn, ID ORDER BY msn DESC)=1)
    ORDER BY ID, CREATED_AT ),

cte3 as (
SELECT *
FROM mr
-- some other operations
)

解决这个问题的理想方法是什么？例如创建常规视图、物化视图或临时表等。我尝试创建视图但出现错误，不确定是否支持。如何在其他后续 CTE 中使用 MATCH_RECOGNIZE 的结果？

当我添加以下内容时，它会出现此错误：

位置 0 处的语法错误行 xx 意外“创建”。

create view filtered_idents AS

SELECT *
FROM cte2
MATCH_RECOGNIZE (

)

【问题讨论】：

Using “match_recognize” in a Common Table Expression in Snowflake
我已经检查了这篇文章。但是，在 MATCH_RECOGNIZE 之后我确实还有 5 个 CTE，所以不要认为使用子查询会更容易和更有效。有没有其他选择？

标签： sql snowflake-cloud-data-platform common-table-expression sql-view match-recognize

【解决方案1】：

这似乎是一个未记录的限制（我要求我们出色的文档团队解决此问题）。

同时，我可以建议将过程划分为步骤以使用match_recognize 结果。

重现错误：

with data as (
    select $1 company, $2 price_date, $3 price
    from values('a',1,10), ('a',2,15)
), cte as (

    select *
    from data match_recognize(
        partition by company
        order by price_date
        measures match_number() as "MATCH_NUMBER"
        all rows per match omit empty matches
        pattern(overavg*)
        define
            overavg as price > avg(price) over (rows between unbounded
                                      preceding and unbounded following)
    )
) 
select * from cte

-- 002362 (0A000): SQL compilation error: MATCH_RECOGNIZE not supported in this context.

两步解决方案：

with data as (
    select $1 company, $2 price_date, $3 price
    from values('a',1,10), ('a',2,15)
)

select *
from data match_recognize(
    partition by company
    order by price_date
    measures match_number() as "MATCH_NUMBER"
    all rows per match omit empty matches
    pattern(overavg*)
    define
        overavg as price > avg(price) over (rows between unbounded
                                  preceding and unbounded following)
)
;


with previous_results as (
    select *
    from table(result_scan(last_query_id()))
)

select * 
from previous_results
;

【讨论】：

一个不错的解决方法。能否以某种方式将其包装在一个视图中（现在我们正在对两个单独的查询进行操作）？理想情况下，我希望有一天会解除这个限制。

【解决方案2】：

Kimi，试试我的 sn-p：

SQL 编译错误：位置 0 的第 11 行语法错误意外“先生”。第 17 行语法错误，位置 6 出现意外的“MEASURES”。

第 9 行似乎缺少一个终止逗号。当我添加一个然后用一个简单的 select 语句完成整个过程时，我不会再遇到语法错误，我只会遇到名称查找错误（当然是预期的）。

【讨论】：