TDengine 选择函数 Max() 用户手册

MAX 函数用户手册

函数定义

MAX(expr)

功能说明

MAX() 函数是一个聚合函数，用于返回指定列或表达式在结果集中的最大值。该函数会忽略 NULL 值，只在所有值都为 NULL 时返回 NULL。

版本要求

• 最低版本: v3.0.0.0

返回值

• 数据类型: 与输入参数相同的数据类型
• 返回内容: 结果集中的最大值

参数说明

参数	类型	说明	取值范围
expr	表达式	要统计的字段表达式	数值类型、字符串类型、时间戳类型

适用数据类型

根据源码分析，MAX 函数支持以下数据类型：

• 数值类型: TINYINT、SMALLINT、INT、BIGINT、FLOAT、DOUBLE、UTINYINT、USMALLINT、UINT、UBIGINT、DECIMAL
• 字符串类型: VARCHAR、NCHAR
• 时间戳类型: TIMESTAMP
• NULL 类型: 支持，但会被忽略

适用范围

• 表类型: 表和超级表
• 查询支持: 支持聚合查询、窗口查询（INTERVAL）
• 函数特性: 聚合函数、选择函数、忽略 NULL 值、支持 TSMA

基本用法示例

单列查询

-- 查找电流的最大值
SELECT MAX(current) FROM meters;

-- 查找电压的最大值
SELECT MAX(voltage) FROM meters;

-- 查找相位的最大值
SELECT MAX(phase) FROM meters;

多列查询

-- 查找多个字段的最大值
SELECT 
    MAX(current) as max_current,
    MAX(voltage) as max_voltage,
    MAX(phase) as max_phase
FROM meters;

字符串类型的最大值

-- 查找设备名称的最大值（按字典序）
SELECT MAX(device_name) FROM devices;

时间戳类型的最大值

-- 查找最新的时间戳
SELECT MAX(ts) as latest_timestamp FROM meters;

智能电表场景应用示例

基于智能电表数据库结构：

-- 数据库和表结构
USE test;
-- meters 超级表包含 ts, current, voltage, phase 字段和 location, groupid 标签

场景1：全局最大值查询

-- 查找所有电表中的最大电流值
SELECT MAX(current) as peak_current FROM meters;

-- 查找所有电表中的最大电压值
SELECT MAX(voltage) as peak_voltage FROM meters;

-- 查找最新的记录时间
SELECT MAX(ts) as latest_record_time FROM meters;

场景2：按区域分组的最大值

-- 查找每个区域的最大电流
SELECT 
    location,
    MAX(current) as max_current,
    MAX(voltage) as max_voltage
FROM meters 
GROUP BY location;

场景3：按设备组分析

-- 查找每个设备组的峰值数据
SELECT 
    groupid,
    location,
    MAX(current) as peak_current,
    MAX(voltage) as peak_voltage,
    MAX(phase) as max_phase,
    COUNT(*) as record_count
FROM meters 
GROUP BY groupid, location
ORDER BY peak_current DESC;

场景4：负载峰值分析

-- 分析各电表的负载峰值
SELECT 
    tbname,
    location,
    MAX(current) as peak_load,
    MAX(voltage) as max_voltage,
    CASE 
        WHEN MAX(current) > 25 THEN '高负载设备'
        WHEN MAX(current) > 15 THEN '中负载设备'
        ELSE '低负载设备'
    END as load_category
FROM meters 
GROUP BY tbname, location
ORDER BY peak_load DESC;

场景5：时间窗口内的峰值

-- 每小时的峰值电流
SELECT 
    location,
    MAX(current) as hourly_peak_current,
    MAX(voltage) as hourly_peak_voltage
FROM meters 
INTERVAL(1h) GROUP BY location;

-- 每天的峰值数据
SELECT 
    location,
    MAX(current) as daily_peak_current,
    MAX(voltage) as daily_peak_voltage,
    MAX(ts) as last_record_time
FROM meters 
INTERVAL(1d) GROUP BY location;

场景6：电网负载监控

-- 监控电网各区域的负载峰值
SELECT 
    location,
    MAX(current) as region_peak_load,
    COUNT(*) as device_count,
    CASE 
        WHEN MAX(current) > 30 THEN '超负荷区域'
        WHEN MAX(current) > 20 THEN '高负荷区域'
        WHEN MAX(current) > 10 THEN '正常负荷区域'
        ELSE '低负荷区域'
    END as load_status
FROM meters 
GROUP BY location
ORDER BY region_peak_load DESC;

场景7：设备容量评估

-- 评估设备的最大使用容量
SELECT 
    tbname,
    location,
    MAX(current) as max_current_usage,
    MAX(voltage) as max_voltage_reading,
    CASE 
        WHEN MAX(current) > 20 THEN '接近容量上限'
        WHEN MAX(current) > 15 THEN '中等使用率'
        ELSE '使用率较低'
    END as capacity_assessment
FROM meters 
GROUP BY tbname, location
HAVING MAX(current) > 10  -- 只显示有一定负载的设备
ORDER BY max_current_usage DESC;

场景8：异常检测

-- 检测电压异常的设备
SELECT 
    location,
    COUNT(*) as total_devices,
    MAX(voltage) as max_voltage,
    CASE 
        WHEN MAX(voltage) > 250 THEN '存在过压风险'
        WHEN MAX(voltage) < 200 THEN '存在欠压问题'
        ELSE '电压正常'
    END as voltage_status
FROM meters 
GROUP BY location;

场景9：历史峰值对比

-- 对比不同时间段的峰值
SELECT 
    location,
    MAX(current) as all_time_peak,
    MAX(CASE WHEN ts >= NOW() - 7d THEN current END) as week_peak,
    MAX(CASE WHEN ts >= NOW() - 1d THEN current END) as today_peak
FROM meters 
GROUP BY location
ORDER BY all_time_peak DESC;

场景10：综合性能指标

-- 设备综合性能评估
SELECT 
    tbname,
    location,
    MAX(current) as peak_current,
    MAX(voltage) as peak_voltage,
    MAX(phase) as max_phase,
    COUNT(*) as data_points,
    MAX(ts) as last_update,
    CASE 
        WHEN MAX(current) > 25 OR MAX(voltage) > 240 THEN '需要关注'
        WHEN MAX(current) > 15 AND MAX(voltage) BETWEEN 200 AND 240 THEN '运行良好'
        ELSE '运行正常'
    END as performance_status
FROM meters 
GROUP BY tbname, location
ORDER BY peak_current DESC, peak_voltage DESC;

与其他聚合函数组合使用

-- 同时查询最大值、最小值、平均值和计数
SELECT 
    location,
    MAX(current) as max_current,
    MIN(current) as min_current,
    AVG(current) as avg_current,
    COUNT(*) as record_count
FROM meters 
GROUP BY location;

-- 分析电流的分布情况
SELECT 
    location,
    MAX(current) as peak_current,
    MIN(current) as min_current,
    MAX(current) - MIN(current) as current_range,
    CASE 
        WHEN MAX(current) - MIN(current) > 15 THEN '电流波动大'
        WHEN MAX(current) - MIN(current) > 5 THEN '电流波动适中'
        ELSE '电流稳定'
    END as stability_assessment
FROM meters 
GROUP BY location;

注意事项

1. NULL 值处理: MAX 函数自动忽略 NULL 值，只在所有值都为 NULL 时返回 NULL
2. 字符串比较: 对于字符串类型，按字典序进行比较
3. 数据类型一致性: 返回值的数据类型与输入参数保持一致
4. 时间戳处理: 对于时间戳类型，返回最新（最大）的时间戳值
5. 性能优化: 在有索引的列上使用 MAX 函数可以获得更好的性能
6. 精度保持: 对于 DECIMAL 类型，返回值保持原有的精度和标度

相关函数

• MIN(): 返回最小值
• AVG(): 返回平均值
• SUM(): 返回总和
• COUNT(): 返回计数
• TOP(): 返回最大的 k 个值
• LAST(): 返回最后的非 NULL 值

关于 TDengine

TDengine 专为物联网IoT平台、工业大数据平台设计。其中，TDengine TSDB 是一款高性能、分布式的时序数据库（Time Series Database），同时它还带有内建的缓存、流式计算、数据订阅等系统功能；TDengine IDMP 是一款AI原生工业数据管理平台，它通过树状层次结构建立数据目录，对数据进行标准化、情景化，并通过 AI 提供实时分析、可视化、事件管理与报警等功能。