如何将过滤器应用于Snowflake的窗口功能？ 我有这样的桌子 transaction_date booked_date 数量 2024-02-10 2024-02-09 50 2024-02-10 2024-02-10 50 2024-02-10 2024-02-11 50 2024-02-11 2024-02-10 50 2024-02-11 2024-02-11 ...

Question

transaction_date

Booked_date

mount2024-02-102024-02-09502024-02-102024-02-10502024-02-102024-02-11502024-02-112024-02-10502024-02-112024-02-11502024-02-112024-02-12502024-02-122024-02-11502024-02-122024-02-12502024-02-122024-02-1350 transaction_dateBooked_date

		指定交易何时发生

指定何时预订交易（即何时会计师知道发生了）

我想要计算两种不同类型的累积总和：
cumulative_amount
• ：每个日期的所有交易的累计量 cumulative_amount_known_as_of_date
：我们知道的所有交易的累积量

这是我期望的结果：

• transaction_date
• cumulative_amount cumulative_amount_known_as_of_date

2024-02-101501002024-02-113002502024-02-12450400

WITH transactions AS (
    SELECT *
    FROM VALUES 
        ('2024-02-10', '2024-02-09', 50),
        ('2024-02-10', '2024-02-10', 50),
        ('2024-02-10', '2024-02-11', 50),
        ('2024-02-11', '2024-02-10', 50),
        ('2024-02-11', '2024-02-11', 50),
        ('2024-02-11', '2024-02-12', 50),
        ('2024-02-12', '2024-02-11', 50),
        ('2024-02-12', '2024-02-12', 50),
        ('2024-02-12', '2024-02-13', 50)
    AS t(TRANSACTION_DATE, BOOKED_DATE, AMOUNT)
)
SELECT
    TRANSACTION_DATE, 
    SUM(SUM(AMOUNT)) 
        OVER (ORDER BY TRANSACTION_DATE) AS cumulative_amount,
    SUM(SUM(IFF(BOOKED_DATE <= TRANSACTION_DATE, AMOUNT, 0))) 
        OVER (ORDER BY TRANSACTION_DATE) AS cumulative_amount_known_as_of_date
FROM TRANSACTIONS
GROUP BY TRANSACTION_DATE

cumulative_amount_as_of_date('2024-02-10', '2024-02-11', 50)我可以正确执行此操作？

		我正在尝试编写雪花查询以获得此结果。这是我的最初尝试：
但	返回不正确的结果。例如，在2/11上，它应该排除行	，因为booked_date大于trassaction_date，但是只要预订日期小于2/11。

问题是这些是字符串而不是日期。 尝试一下：

WITH transactions AS (
    SELECT *
    FROM VALUES 
        (CAST('2024-02-10' AS DATE), CAST('2024-02-09' AS DATE), 50),
        (CAST('2024-02-10' AS DATE), CAST('2024-02-10' AS DATE), 50),
        (CAST('2024-02-10' AS DATE), CAST('2024-02-11' AS DATE), 50),
        (CAST('2024-02-11' AS DATE), CAST('2024-02-10' AS DATE), 50),
        (CAST('2024-02-11' AS DATE), CAST('2024-02-11' AS DATE), 50),
        (CAST('2024-02-11' AS DATE), CAST('2024-02-12' AS DATE), 50),
        (CAST('2024-02-12' AS DATE), CAST('2024-02-11' AS DATE), 50),
        (CAST('2024-02-12' AS DATE), CAST('2024-02-12' AS DATE), 50),
        (CAST('2024-02-12' AS DATE), CAST('2024-02-13' AS DATE), 50)
    AS t(TRANSACTION_DATE, BOOKED_DATE, AMOUNT)
)
SELECT
    TRANSACTION_DATE, 
    SUM(SUM(AMOUNT)) 
        OVER (ORDER BY TRANSACTION_DATE) AS cumulative_amount,
    SUM(SUM(IFF(BOOKED_DATE <= TRANSACTION_DATE, AMOUNT, 0))) 
        OVER (ORDER BY TRANSACTION_DATE) AS cumulative_amount_known_as_of_date
FROM TRANSACTIONS
GROUP BY TRANSACTION_DATE

注意某些版本的SQL不会具有日期类型，您应该使用DateTime。 我相信雪花有约会