DolphinDB SQL查询：从简单到复杂

目录

• • 摘要
  • 一、基础查询
  • • [1.1 SELECT语句](#1.1 SELECT语句)
    • [1.2 条件过滤](#1.2 条件过滤)
    • [1.3 排序与限制](#1.3 排序与限制)
  • 二、聚合查询
  • • [2.1 基本聚合](#2.1 基本聚合)
    • [2.2 分组聚合](#2.2 分组聚合)
    • [2.3 分组集](#2.3 分组集)
  • 三、连接查询
  • • [3.1 连接类型](#3.1 连接类型)
    • [3.2 内连接](#3.2 内连接)
    • [3.3 左连接](#3.3 左连接)
    • [3.4 多表连接](#3.4 多表连接)
  • 四、子查询
  • • [4.1 标量子查询](#4.1 标量子查询)
    • [4.2 行子查询](#4.2 行子查询)
    • [4.3 表子查询](#4.3 表子查询)
  • 五、窗口函数
  • • [5.1 排序函数](#5.1 排序函数)
    • [5.2 聚合窗口函数](#5.2 聚合窗口函数)
    • [5.3 偏移函数](#5.3 偏移函数)
  • 六、时间序列查询
  • • [6.1 时间窗口](#6.1 时间窗口)
    • [6.2 时间对齐](#6.2 时间对齐)
    • [6.3 重采样](#6.3 重采样)
  • 七、复杂查询实战
  • • [7.1 分组Top-N](#7.1 分组Top-N)
    • [7.2 同比环比](#7.2 同比环比)
    • [7.3 连续区间](#7.3 连续区间)
  • 八、查询优化
  • • [8.1 执行计划](#8.1 执行计划)
    • [8.2 优化建议](#8.2 优化建议)
    • [8.3 常见问题](#8.3 常见问题)
  • 九、总结
  • 参考资料

摘要

本文系统介绍DolphinDB SQL查询语言。从基础SELECT语句到复杂的多表关联，从条件过滤到分组聚合，从子查询到窗口函数，逐步带领读者掌握DolphinDB SQL查询的核心技能。通过丰富的实战案例，帮助读者编写高效的数据查询语句。

---

一、基础查询

1.1 SELECT语句

// 创建示例表
t = table(
    1..10 as id,
    `A`B`C`A`B`C`A`B`C`A as category,
    10 20 30 15 25 35 12 22 32 18 as value,
    2024.01.01 + 0..9 as date
)

// 查询所有列
select * from t

// 查询指定列
select id, category, value from t

// 列别名
select id as device_id, value as temperature from t

// 计算列
select id, value, value * 1.8 + 32 as fahrenheit from t

1.2 条件过滤

// WHERE条件
select * from t where id > 5

// 多条件
select * from t where id > 3 and value < 30

// IN条件
select * from t where category in [`A, `B]

// BETWEEN条件
select * from t where value between 15 and 30

// LIKE模糊匹配
select * from t where category like "A%"

// NULL判断
select * from t where value is not NULL

1.3 排序与限制

// 排序
select * from t order by value desc
select * from t order by category, value desc

// 限制行数
select top 5 * from t
select top 5 * from t order by value desc

// 分页
select * from t limit 5 offset 3  // 从第4行开始取5行

---

二、聚合查询

2.1 基本聚合

// 聚合函数
select count(*) as cnt from t
select sum(value) as total from t
select avg(value) as average from t
select max(value) as max_val, min(value) as min_val from t

// 多聚合
select count(*) as cnt,
       sum(value) as total,
       avg(value) as average,
       std(value) as std_dev
from t

2.2 分组聚合

// GROUP BY
select category, 
       count(*) as cnt,
       sum(value) as total,
       avg(value) as average
from t
group by category

// 多列分组
select category, 
       date,
       count(*) as cnt,
       avg(value) as avg_val
from t
group by category, date

// HAVING过滤
select category, avg(value) as avg_val
from t
group by category
having avg(value) > 20

2.3 分组集

// ROLLUP
select category, date, sum(value) as total
from t
group by rollup(category, date)

// CUBE
select category, date, sum(value) as total
from t
group by cube(category, date)

// GROUPING SETS
select category, date, sum(value) as total
from t
group by grouping sets((category), (date), (category, date))

---

三、连接查询

3.1 连接类型

连接类型
INNER JOIN

内连接
交集
LEFT JOIN

左连接
左表全部
RIGHT JOIN

右连接
右表全部
FULL JOIN

全连接
并集

3.2 内连接

// 创建示例表
t1 = table(1..5 as id, `A`B`C`D`E as name)
t2 = table(1..3 as id, 100..102 as value)

// 内连接
select * from lj(t1, t2, `id)
// 或
select t1.id, t1.name, t2.value 
from t1
inner join t2 on t1.id = t2.id

3.3 左连接

// 左连接
select * from lj(t1, t2, `id)

/*
id name value
1  A    100
2  B    101
3  C    102
4  D    NULL
5  E    NULL
*/

3.4 多表连接

// 多表连接
t1 = table(1..5 as id, `A`B`C`D`E as name)
t2 = table(1..5 as id, 100..104 as value)
t3 = table(1..5 as id, `X`Y`Z`X`Y as type)

select t1.id, t1.name, t2.value, t3.type
from t1
left join t2 on t1.id = t2.id
left join t3 on t1.id = t3.id

---

四、子查询

4.1 标量子查询

// 标量子查询（返回单个值）
select * from t 
where value > (select avg(value) from t)

// 在SELECT中使用
select id, value, 
       (select avg(value) from t) as avg_value
from t

4.2 行子查询

// 行子查询（返回多行）
select * from t
where category in (select distinct category from t where value > 25)

// EXISTS子查询
select * from t1
where exists (select * from t2 where t2.id = t1.id)

4.3 表子查询

// 表子查询（返回表）
select category, avg_val
from (
    select category, avg(value) as avg_val
    from t
    group by category
)
where avg_val > 20

// WITH子句（CTE）
with 
stats as (
    select category, avg(value) as avg_val
    from t
    group by category
)
select * from stats where avg_val > 20

---

五、窗口函数

5.1 排序函数

// ROW_NUMBER
select id, category, value,
       row_number() over (partition by category order by value desc) as rank
from t

// RANK（有并列）
select id, category, value,
       rank() over (order by value desc) as rank
from t

// DENSE_RANK（连续排名）
select id, category, value,
       dense_rank() over (order by value desc) as rank
from t

5.2 聚合窗口函数

// 累积聚合
select id, value,
       sum(value) over (order by id) as cumsum,
       avg(value) over (order by id rows between 2 preceding and current row) as mavg
from t

// 分组窗口
select id, category, value,
       sum(value) over (partition by category order by id) as category_cumsum
from t

5.3 偏移函数

// LAG/LEAD
select id, value,
       lag(value, 1) over (order by id) as prev_value,
       lead(value, 1) over (order by id) as next_value
from t

// FIRST/LAST
select id, value,
       first(value) over (order by id) as first_val,
       last(value) over (order by id) as last_val
from t

---

六、时间序列查询

6.1 时间窗口

// 创建时间序列数据
t = table(
    1..100 as id,
    2024.01.01T00:00:00 + 0..99 * 60000 as timestamp,  // 每分钟一条
    rand(100.0, 100) as value
)

// 时间窗口聚合
select bar(timestamp, 10m) as time_window,
       avg(value) as avg_val,
       max(value) as max_val,
       count(*) as cnt
from t
group by bar(timestamp, 10m)

6.2 时间对齐

// 时间对齐
select * from t 
where timestamp between 2024.01.01T00:00:00 and 2024.01.01T01:00:00

// 按小时聚合
select bar(timestamp, 1h) as hour,
       avg(value) as avg_val
from t
group by bar(timestamp, 1h)

6.3 重采样

// 重采样（1分钟→5分钟）
select bar(timestamp, 5m) as time_5m,
       first(value) as open,
       max(value) as high,
       min(value) as low,
       last(value) as close
from t
group by bar(timestamp, 5m)

---

七、复杂查询实战

7.1 分组Top-N

// 每个类别取前3条
select * from (
    select id, category, value,
           row_number() over (partition by category order by value desc) as rank
    from t
) where rank <= 3

7.2 同比环比

// 环比计算
select date, value,
       lag(value, 1) over (order by date) as prev_value,
       (value - lag(value, 1) over (order by date)) / lag(value, 1) over (order by date) as mom_rate
from t

// 同比计算
select date, value,
       lag(value, 12) over (order by date) as prev_year_value,
       (value - lag(value, 12) over (order by date)) / lag(value, 12) over (order by date) as yoy_rate
from t

7.3 连续区间

// 查找连续值
select * from (
    select id, value,
           id - row_number() over (order by id) as grp
    from t
    where value > 20
) 
group by grp
having count(*) >= 3  // 连续3个以上

---

八、查询优化

8.1 执行计划

// 查看执行计划
explain select * from t where id > 5

// 分析查询性能
timer select count(*) from t

8.2 优化建议

优化项	说明
分区裁剪	在分区列上过滤
索引使用	在索引列上查询
减少扫描	只查询需要的列
避免全表	使用WHERE条件

8.3 常见问题

// 避免SELECT *
select id, value from t  // 好
select * from t          // 避免

// 使用分区过滤
select * from t 
where date between 2024.01.01 and 2024.01.31  // 好

// 避免函数包装索引列
select * from t where date = 2024.01.15      // 好
select * from t where year(date) = 2024      // 避免

---

九、总结

本文系统介绍了DolphinDB SQL查询：

1. 基础查询：SELECT、WHERE、ORDER BY
2. 聚合查询：GROUP BY、HAVING、分组集
3. 连接查询：内连接、左连接、多表连接
4. 子查询：标量、行、表子查询
5. 窗口函数：排序、聚合、偏移
6. 时间序列：时间窗口、重采样
7. 查询优化：执行计划、优化建议

思考题：

1. 如何选择合适的连接类型？
2. 窗口函数和GROUP BY有什么区别？
3. 如何优化大数据量查询？

---

参考资料

• DolphinDB SQL查询
• DolphinDB窗口函数