如何为postgresql中查询的前n行分配唯一标识符？答案

【问题标题】：How to assign a unique identifier to the first n rows of a query in postgresql?如何为postgresql中查询的前n行分配唯一标识符？
【发布时间】：2022-01-27 01:36:09
【问题描述】：

在 postgres SQL 中，如何将唯一标识符 (rn?) 分配给查询的前 n 行，并将唯一标识符分配给随后的 n 行？目标是按代码和唯一标识符进行分组，有关详细信息，请参阅实际与预期的屏幕截图。

实际查询

SELECT 
    *,
    SUM(eps_diluted) OVER (PARTITION BY ticker ORDER BY rn ASC ROWS BETWEEN CURRENT ROW AND 3 FOLLOWING) AS test_eps_diluted
FROM(
    SELECT 
        "PK",
        ticker,
        period_end_date,
        eps_diluted,
        ROW_NUMBER() OVER ( PARTITION BY ticker ORDER BY period_end_date DESC ) rn
    FROM "ANALYTICS"."vQUARTERLY_MASTER_MATERIALIZED" 
    --WHERE ticker = 'ACN' 
    ORDER BY ticker, period_end_date DESC
    ) q
ORDER BY ticker, period_end_date desc

Actual result

Expected result

【问题讨论】：

标签： postgresql group-by row-number

【解决方案1】：

我找到了一种使用模数的解决方法。此解决方案允许我对最近几个季度、随后的 4 个季度等的 eps_diluted 求和。详见解决方案截图。

SELECT 
    * 
FROM (
    SELECT 
        *,
        SUM(eps_diluted) OVER (PARTITION BY ticker ORDER BY rn ASC ROWS BETWEEN CURRENT ROW AND 3 FOLLOWING) AS sum_eps_diluted,
        1+ ((rn - 1) % 4) AS modulo
    FROM(
        SELECT 
            "PK",
            ticker,
            period_end_date,
            eps_diluted,
            ROW_NUMBER() OVER ( PARTITION BY ticker ORDER BY period_end_date DESC ROWS BETWEEN CURRENT ROW AND 3 FOLLOWING) rn
        FROM "ANALYTICS"."vQUARTERLY_MASTER_MATERIALIZED" 
        ORDER BY ticker, period_end_date DESC
        ) t1
    --WHERE 1+ ((rn - 1) % 4) = 1
    ORDER BY ticker, period_end_date desc ) t2
WHERE modulo = 1

Solution

【讨论】：