查询某时间范围内数据，按天统计并补齐空白日期为0的通用实现方案

nbsaas-boot

于 2025-07-18 01:45:00 发布

阅读量405

点赞数 4

CC 4.0 BY-SA版权

文章标签：数据库 sql oracle

本文链接：https://blog.csdn.net/u010362741/article/details/149424639

在实际的数据分析和报表需求中，我们经常会遇到以下场景：

“我要统计过去 7 天（或任意时间段）内每天的销售数据（或用户注册数、访问次数等），即使某些天没有数据，也要补0，保证图表或报表整齐。”

这是一个非常常见的需求，但初学者常常因为“日期缺失”而无法得到连续的结果。本篇文章将介绍如何在不同的数据库系统中实现这种“时间维度补全+聚合统计”的能力。

📌 场景说明

假设我们有如下表结构：

orders (
  id BIGINT PRIMARY KEY,
  order_time DATETIME,
  amount DECIMAL
)

目标：

按天统计订单数量
时间范围：如 2024-01-01 ~ 2024-01-07
若某天没有订单，也需要返回该天，订单数为 0

🧠 思路拆解

实现这个功能，一般分三步：

构造一个时间序列（时间维度）
通常用递归、函数或维度表生成。
按天聚合业务数据
比如 SELECT DATE(order_time), COUNT(*) FROM orders GROUP BY DATE(order_time)
左连接时间序列与聚合结果，使用 ISNULL 或 COALESCE 补0

🔧 各类数据库实现方案

✅ 1. SQL Server（使用递归CTE）

DECLARE @startDate DATE = '2024-01-01';
DECLARE @endDate DATE = '2024-01-07';

WITH DateSeries AS (
    SELECT @startDate AS dt
    UNION ALL
    SELECT DATEADD(DAY, 1, dt)
    FROM DateSeries
    WHERE dt < @endDate
),
OrderStats AS (
    SELECT CAST(order_time AS DATE) AS order_date, COUNT(*) AS cnt
    FROM orders
    WHERE order_time BETWEEN @startDate AND @endDate
    GROUP BY CAST(order_time AS DATE)
)
SELECT d.dt, ISNULL(o.cnt, 0) AS order_count
FROM DateSeries d
LEFT JOIN OrderStats o ON d.dt = o.order_date
ORDER BY d.dt
OPTION (MAXRECURSION 1000);

✅ 2. MySQL 8.0+（使用递归CTE）

WITH RECURSIVE DateSeries AS (
    SELECT DATE('2024-01-01') AS dt
    UNION ALL
    SELECT DATE_ADD(dt, INTERVAL 1 DAY)
    FROM DateSeries
    WHERE dt < '2024-01-07'
),
OrderStats AS (
    SELECT DATE(order_time) AS order_date, COUNT(*) AS cnt
    FROM orders
    WHERE order_time BETWEEN '2024-01-01' AND '2024-01-07'
    GROUP BY DATE(order_time)
)
SELECT d.dt, IFNULL(o.cnt, 0) AS order_count
FROM DateSeries d
LEFT JOIN OrderStats o ON d.dt = o.order_date
ORDER BY d.dt;

✅ 3. PostgreSQL（使用 `generate_series`）

WITH DateSeries AS (
    SELECT generate_series('2024-01-01'::DATE, '2024-01-07'::DATE, INTERVAL '1 day') AS dt
),
OrderStats AS (
    SELECT order_time::DATE AS order_date, COUNT(*) AS cnt
    FROM orders
    WHERE order_time BETWEEN '2024-01-01' AND '2024-01-07'
    GROUP BY order_time::DATE
)
SELECT d.dt::DATE, COALESCE(o.cnt, 0) AS order_count
FROM DateSeries d
LEFT JOIN OrderStats o ON d.dt = o.order_date
ORDER BY d.dt;

✅ 4. Oracle（使用 CONNECT BY）

SELECT 
    TO_DATE('2024-01-01', 'YYYY-MM-DD') + LEVEL - 1 AS dt,
    NVL(o.cnt, 0) AS order_count
FROM dual
CONNECT BY LEVEL <= (TO_DATE('2024-01-07', 'YYYY-MM-DD') - TO_DATE('2024-01-01', 'YYYY-MM-DD')) + 1
LEFT JOIN (
    SELECT TRUNC(order_time) AS order_date, COUNT(*) AS cnt
    FROM orders
    WHERE order_time BETWEEN TO_DATE('2024-01-01', 'YYYY-MM-DD') AND TO_DATE('2024-01-07', 'YYYY-MM-DD')
    GROUP BY TRUNC(order_time)
) o
ON TO_DATE('2024-01-01', 'YYYY-MM-DD') + LEVEL - 1 = o.order_date
ORDER BY dt;

✅ 5. SQLite（不支持递归，构造数字序列）

-- 构造连续数字
WITH RECURSIVE nums(n) AS (
    SELECT 0
    UNION ALL
    SELECT n + 1 FROM nums WHERE n < 6
),
DateSeries AS (
    SELECT DATE('2024-01-01', '+' || n || ' days') AS dt FROM nums
),
OrderStats AS (
    SELECT DATE(order_time) AS order_date, COUNT(*) AS cnt
    FROM orders
    WHERE order_time BETWEEN '2024-01-01' AND '2024-01-07'
    GROUP BY DATE(order_time)
)
SELECT d.dt, IFNULL(o.cnt, 0) AS order_count
FROM DateSeries d
LEFT JOIN OrderStats o ON d.dt = o.order_date
ORDER BY d.dt;