doris sql 多个元素分为同一组

doris 表中有如下字段:

sql 复制代码

create table IF NOT EXISTS  cdn.edge_client_area_req_fivemin
(
  `time` DATETIME  NOT NULL COMMENT "时间",
    `metric` VARCHAR(50) NOT NULL COMMENT "指标名称",
    `cp` VARCHAR(100) COMMENT "公司",
    `productId` INT COMMENT "产品id，7,超低时延",
    `provider` VARCHAR(20)  COMMENT "平面",
    `clientArea` VARCHAR(50) COMMENT "客户端省份",
    `domain` VARCHAR(100) COMMENT "域名",
    `clientOperator` VARCHAR(20) COMMENT "客户端运营商",
    `ipProtocol` VARCHAR(20) COMMENT "ip协议: 0：ipv4, 1:ipv6",
    `stateCode` INT COMMENT "状态码",
    `orderId` VARCHAR(100) COMMENT "订购号",
    `sourceCode` VARCHAR(100) COMMENT "来源编码：01:边缘节点 03:HCDH",
    `value` DOUBLE COMMENT "请求数"
)

现在想根据productId得值有(1,2,3,4),现在想将1,2并为一组，3，4并为一组,5，6并为一组进行分组，如何写sql

方案一：CASE WHEN（最清晰、最推荐）

sql 复制代码

SELECT
    CASE
        WHEN productId IN (1, 2) THEN 'group_1_2'
        WHEN productId IN (3, 4) THEN 'group_3_4'
        WHEN productId IN (5, 6) THEN 'group_5_6'
        ELSE 'other'
    END AS product_group,
    SUM(value) AS total_requests
FROM cdn.edge_client_area_req_fivemin
WHERE productId IN (1, 2, 3, 4, 5, 6)
GROUP BY
    CASE
        WHEN productId IN (1, 2) THEN 'group_1_2'
        WHEN productId IN (3, 4) THEN 'group_3_4'
        WHEN productId IN (5, 6) THEN 'group_5_6'
        ELSE 'other'
    END;

方案二：IF嵌套（写法更紧凑）

sql 复制代码

SELECT
    IF(productId IN (1, 2), 'group_1_2',
       IF(productId IN (3, 4), 'group_3_4',
          IF(productId IN (5, 6), 'group_5_6', 'other'))
    ) AS product_group,
    SUM(value) AS total_requests
FROM cdn.edge_client_area_req_fivemin
WHERE productId IN (1, 2, 3, 4, 5, 6)
GROUP BY product_group;

方案三：结合时间 / 维度统计（真实业务常用）

sql 复制代码

SELECT
    time,
    CASE
        WHEN productId IN (1, 2) THEN 'group_1_2'
        WHEN productId IN (3, 4) THEN 'group_3_4'
        WHEN productId IN (5, 6) THEN 'group_5_6'
    END AS product_group,
    clientArea,
    SUM(value) AS total_requests
FROM cdn.edge_client_area_req_fivemin
WHERE productId IN (1, 2, 3, 4, 5, 6)
GROUP BY time, product_group, clientArea;

方案四：使用 CTE（逻辑更清晰，便于维护）

sql 复制代码

WITH grouped_data AS (
    SELECT
        *,
        CASE
            WHEN productId IN (1, 2) THEN 'group_1_2'
            WHEN productId IN (3, 4) THEN 'group_3_4'
            WHEN productId IN (5, 6) THEN 'group_5_6'
        END AS product_group
    FROM cdn.edge_client_area_req_fivemin
    WHERE productId IN (1, 2, 3, 4, 5, 6)
)
SELECT
    product_group,
    clientOperator,
    ipProtocol,
    SUM(value) AS total_requests
FROM grouped_data
GROUP BY product_group, clientOperator, ipProtocol;