达梦(DM8)基于 LBS（位置服务）教程

一、环境准备

1.1 达梦数据库版本要求

• 达梦数据库 DM8 及以上版本

• 需安装空间数据组件（可选）

1.2 创建测试数据库

-- 创建表空间
CREATE TABLESPACE lbs_data DATAFILE 'lbs_data.dbf' SIZE 1024;

-- 创建用户
CREATE USER lbs_user IDENTIFIED BY "Lbs_123456"
DEFAULT TABLESPACE lbs_data;

GRANT RESOURCE, VTI TO lbs_user;

---

二、空间数据类型基础

2.1 达梦支持的空间数据类型

-- 1. 点（POINT）
CREATE TABLE location_points (
    id INT PRIMARY KEY,
    name VARCHAR(100),
    -- 经度,纬度格式
    coordinate ST_POINT
);

-- 2. 线（LINESTRING）
CREATE TABLE road_segments (
    id INT PRIMARY KEY,
    road_name VARCHAR(100),
    path ST_LINESTRING
);

-- 3. 多边形（POLYGON）
CREATE TABLE service_areas (
    id INT PRIMARY KEY,
    area_name VARCHAR(100),
    boundary ST_POLYGON
);

-- 4. 几何集合（GEOMETRYCOLLECTION）
CREATE TABLE mixed_features (
    id INT PRIMARY KEY,
    features ST_GEOMETRY
);

---

三、LBS数据建模

3.1 创建地理位置表

-- 创建店铺位置表
CREATE TABLE store_locations (
    store_id INT PRIMARY KEY,
    store_name VARCHAR(200),
    address VARCHAR(500),
    -- 存储经纬度，SRID 4326表示WGS84坐标系
    location ST_POINT SRID 4326,
    business_hours VARCHAR(100),
    phone VARCHAR(20),
    create_time TIMESTAMP DEFAULT CURRENT_TIMESTAMP()
);

-- 创建空间索引
CREATE SPATIAL INDEX idx_store_location 
ON store_locations(location) 
STORAGE(ON MAIN);

-- 创建用户位置记录表
CREATE TABLE user_locations (
    record_id BIGINT IDENTITY(1,1) PRIMARY KEY,
    user_id INT,
    device_id VARCHAR(50),
    location ST_POINT SRID 4326,
    accuracy FLOAT,
    speed FLOAT,
    record_time TIMESTAMP DEFAULT CURRENT_TIMESTAMP(),
    address VARCHAR(500)
);

CREATE INDEX idx_user_time ON user_locations(user_id, record_time);
CREATE SPATIAL INDEX idx_user_location ON user_locations(location);

3.2 创建地理围栏表

-- 地理围栏表
CREATE TABLE geo_fences (
    fence_id INT PRIMARY KEY,
    fence_name VARCHAR(200),
    fence_type VARCHAR(20), -- 'CIRCLE', 'POLYGON', 'ROUTE'
    -- 圆形围栏：圆心和半径
    center ST_POINT SRID 4326,
    radius_meters FLOAT,
    -- 多边形围栏
    polygon ST_POLYGON SRID 4326,
    -- 路线围栏（缓冲区）
    route ST_LINESTRING SRID 4326,
    route_buffer_meters FLOAT,
    valid_start DATE,
    valid_end DATE,
    properties VARCHAR(4000) -- JSON格式存储额外属性
);

-- 围栏触发记录表
CREATE TABLE fence_triggers (
    trigger_id BIGINT IDENTITY(1,1) PRIMARY KEY,
    fence_id INT,
    user_id INT,
    event_type VARCHAR(20), -- 'ENTER', 'EXIT', 'INSIDE'
    location ST_POINT SRID 4326,
    trigger_time TIMESTAMP DEFAULT CURRENT_TIMESTAMP(),
    FOREIGN KEY (fence_id) REFERENCES geo_fences(fence_id)
);

---

四、数据操作示例

4.1 插入空间数据

-- 插入点数据（WGS84坐标系）
INSERT INTO store_locations (store_id, store_name, address, location) 
VALUES (
    1, 
    '北京旗舰店', 
    '北京市海淀区中关村大街1号',
    ST_POINT('POINT(116.316836 39.983825)')
);

-- 使用ST_GeomFromText插入
INSERT INTO store_locations (store_id, store_name, location) 
VALUES (
    2,
    '上海中心店',
    ST_GeomFromText('POINT(121.472641 31.231706)', 4326)
);

-- 插入多边形围栏（商业区范围）
INSERT INTO geo_fences (fence_id, fence_name, fence_type, polygon)
VALUES (
    1,
    '中关村科技园区',
    'POLYGON',
    ST_GeomFromText('POLYGON((116.306 39.977, 116.320 39.977, 116.320 39.990, 116.306 39.990, 116.306 39.977))', 4326)
);

4.2 批量导入地理位置数据

-- 创建批量导入存储过程
CREATE OR REPLACE PROCEDURE batch_import_stores()
AS
BEGIN
    -- 示例：批量插入店铺数据
    INSERT INTO store_locations (store_id, store_name, location) 
    VALUES 
    (101, '门店A', ST_Point(116.407526, 39.904030)),
    (102, '门店B', ST_Point(116.403963, 39.915215)),
    (103, '门店C', ST_Point(116.350676, 39.958702)),
    (104, '门店D', ST_Point(116.486124, 39.996786));
    
    COMMIT;
END;
/

CALL batch_import_stores();

---

五、LBS查询示例

5.1 基础空间查询

-- 1. 查找最近的店铺（按距离排序）
SELECT 
    store_id,
    store_name,
    address,
    ST_Distance(
        location,
        ST_Point(116.407526, 39.904030) -- 用户当前位置
    ) * 111319.5 AS distance_meters  -- 转换为米
FROM store_locations
ORDER BY distance_meters
LIMIT 10;

-- 2. 查找指定半径内的店铺（5公里范围内）
SELECT 
    store_id,
    store_name,
    ST_AsText(location) AS coordinates,
    ST_Distance(location, ST_Point(116.407526, 39.904030)) * 111319.5 AS distance_m
FROM store_locations
WHERE ST_Distance(location, ST_Point(116.407526, 39.904030)) * 111319.5 <= 5000
ORDER BY distance_m;

-- 3. 空间范围查询（矩形区域）
SELECT *
FROM store_locations
WHERE ST_Within(location, 
    ST_GeomFromText('POLYGON((116.30 39.90, 116.45 39.90, 116.45 40.00, 116.30 40.00, 116.30 39.90))', 4326)
);

5.2 地理围栏查询

-- 1. 检查用户是否进入围栏
CREATE OR REPLACE PROCEDURE check_fence_trigger(
    p_user_id INT,
    p_longitude FLOAT,
    p_latitude FLOAT
)
AS
    v_point ST_POINT;
    v_fence_id INT;
    v_distance FLOAT;
BEGIN
    -- 创建用户位置点
    v_point := ST_Point(p_longitude, p_latitude);
    
    -- 检查圆形围栏
    FOR fence IN (
        SELECT fence_id, center, radius_meters
        FROM geo_fences 
        WHERE fence_type = 'CIRCLE'
          AND valid_start <= SYSDATE
          AND valid_end >= SYSDATE
    ) LOOP
        v_distance := ST_Distance(v_point, fence.center) * 111319.5;
        
        IF v_distance <= fence.radius_meters THEN
            -- 记录触发事件
            INSERT INTO fence_triggers(fence_id, user_id, event_type, location)
            VALUES (fence.fence_id, p_user_id, 'ENTER', v_point);
            COMMIT;
        END IF;
    END LOOP;
    
    -- 检查多边形围栏
    FOR fence IN (
        SELECT fence_id, polygon
        FROM geo_fences 
        WHERE fence_type = 'POLYGON'
          AND valid_start <= SYSDATE
          AND valid_end >= SYSDATE
    ) LOOP
        IF ST_Within(v_point, fence.polygon) THEN
            INSERT INTO fence_triggers(fence_id, user_id, event_type, location)
            VALUES (fence.fence_id, p_user_id, 'INSIDE', v_point);
            COMMIT;
        END IF;
    END LOOP;
END;
/

-- 2. 查询围栏内所有用户
SELECT DISTINCT u.user_id, u.record_time
FROM user_locations u
JOIN geo_fences g ON ST_Within(u.location, g.polygon)
WHERE g.fence_id = 1
  AND u.record_time >= SYSDATE - INTERVAL '1' HOUR;

5.3 路径规划相关查询

-- 1. 查找沿路径的店铺（路径缓冲区查询）
CREATE OR REPLACE FUNCTION find_stores_along_route(
    route_coords CLOB,  -- JSON格式的路径点
    buffer_distance FLOAT  -- 缓冲区距离（米）
) 
RETURN TABLE(store_id INT, store_name VARCHAR(200), distance FLOAT)
AS
    v_route ST_LINESTRING;
    v_buffer ST_POLYGON;
BEGIN
    -- 将JSON路径转换为线（这里简化处理，实际需要解析JSON）
    v_route := ST_GeomFromText('LINESTRING(116.306 39.977, 116.310 39.978, 116.315 39.980)', 4326);
    
    -- 创建缓冲区（将度转换为近似米）
    v_buffer := ST_Buffer(v_route, buffer_distance / 111319.5);
    
    -- 返回缓冲区内的店铺
    RETURN QUERY
    SELECT s.store_id, s.store_name, 
           ST_Distance(s.location, v_route) * 111319.5 AS distance
    FROM store_locations s
    WHERE ST_Within(s.location, v_buffer)
    ORDER BY distance;
END;
/

-- 2. 最近设施查询（包含路线距离）
WITH user_location AS (
    SELECT ST_Point(116.407526, 39.904030) AS loc
),
store_distances AS (
    SELECT 
        s.*,
        ST_Distance(s.location, u.loc) * 111319.5 AS straight_distance,
        -- 这里可以调用路线规划API计算实际路线距离
        0 AS route_distance  -- 实际应用中替换为计算值
    FROM store_locations s, user_location u
)
SELECT *
FROM store_distances
ORDER BY route_distance
LIMIT 5;

---

六、性能优化

6.1 空间索引优化

-- 1. 查看空间索引信息
SELECT * FROM USER_SDO_GEOM_METADATA;

-- 2. 重建空间索引
ALTER SPATIAL INDEX idx_store_location REBUILD;

-- 3. 创建复合索引
CREATE INDEX idx_store_loc_name ON store_locations(store_name, location);

-- 4. 分区表提升性能
CREATE TABLE user_locations_partitioned (
    record_id BIGINT,
    user_id INT,
    location ST_POINT,
    record_time TIMESTAMP
)
PARTITION BY RANGE(record_time) (
    PARTITION p202401 VALUES LESS THAN ('2024-02-01'),
    PARTITION p202402 VALUES LESS THAN ('2024-03-01'),
    PARTITION p202403 VALUES LESS THAN ('2024-04-01')
);

CREATE SPATIAL INDEX idx_part_loc ON user_locations_partitioned(location) LOCAL;

6.2 查询优化技巧

-- 1. 使用空间索引提示
SELECT /*+ INDEX(store_locations idx_store_location) */ 
    store_id, store_name
FROM store_locations
WHERE ST_DWithin(location, ST_Point(116.407526, 39.904030), 0.05); -- 约5公里

-- 2. 预计算距离并缓存
CREATE TABLE store_distance_cache AS
SELECT 
    s1.store_id AS store_id1,
    s2.store_id AS store_id2,
    ST_Distance(s1.location, s2.location) * 111319.5 AS distance_m
FROM store_locations s1
CROSS JOIN store_locations s2
WHERE s1.store_id < s2.store_id;

CREATE INDEX idx_cache ON store_distance_cache(store_id1, store_id2);

-- 3. 分页查询优化
CREATE OR REPLACE PROCEDURE find_nearby_stores_paged(
    p_longitude FLOAT,
    p_latitude FLOAT,
    p_radius_km FLOAT,
    p_page_num INT,
    p_page_size INT
)
AS
    v_offset INT;
    v_limit INT;
BEGIN
    v_offset := (p_page_num - 1) * p_page_size;
    v_limit := p_page_size;
    
    SELECT * FROM (
        SELECT 
            store_id,
            store_name,
            ST_Distance(location, ST_Point(p_longitude, p_latitude)) * 111319.5 AS distance_m,
            ROW_NUMBER() OVER (ORDER BY distance_m) AS rn
        FROM store_locations
        WHERE ST_Distance(location, ST_Point(p_longitude, p_latitude)) * 111319.5 <= p_radius_km * 1000
    ) WHERE rn > v_offset AND rn <= v_offset + v_limit;
END;
/

---

七、Java应用示例

Spring Boot集成

# application.yml
spring:
  datasource:
    driver-class-name: dm.jdbc.driver.DmDriver
    url: jdbc:dm://localhost:5236/LBS_DB
    username: lbs_user
    password: Lbs_123456
    hikari:
      maximum-pool-size: 20
      minimum-idle: 5

@Repository
public interface StoreRepository {
    @Query(value = """
        SELECT store_id as id, store_name as name, 
               ST_X(location) as lng, ST_Y(location) as lat,
               ST_Distance(location, ST_Point(:lng, :lat)) * 111319.5 as distance
        FROM store_locations 
        WHERE ST_DWithin(location, ST_Point(:lng, :lat), :radius / 111319.5)
        ORDER BY distance
        LIMIT :limit
        """, nativeQuery = true)
    List<StoreProjection> findNearbyStores(
        @Param("lng") double longitude,
        @Param("lat") double latitude,
        @Param("radius") double radiusMeters,
        @Param("limit") int limit
    );
}

@Service
public class LbsService {
    @Autowired
    private StoreRepository storeRepository;
    
    public List<StoreDTO> findStoresWithinRadius(
        LocationPoint center, 
        double radiusKm, 
        int maxResults
    ) {
        return storeRepository.findNearbyStores(
            center.getLongitude(),
            center.getLatitude(),
            radiusKm * 1000,
            maxResults
        ).stream()
        .map(proj -> new StoreDTO(
            proj.getId(),
            proj.getName(),
            proj.getLng(),
            proj.getLat(),
            proj.getDistance()
        ))
        .collect(Collectors.toList());
    }
}

---

八、最佳实践

8.1 数据验证

-- 创建验证约束
ALTER TABLE store_locations 
ADD CONSTRAINT chk_valid_location 
CHECK (
    ST_X(location) BETWEEN -180 AND 180 
    AND ST_Y(location) BETWEEN -90 AND 90
);

-- 创建验证函数
CREATE OR REPLACE FUNCTION validate_geo_fence(
    p_fence_type VARCHAR,
    p_center ST_POINT,
    p_radius FLOAT,
    p_polygon ST_POLYGON
) 
RETURN BOOLEAN
AS
BEGIN
    IF p_fence_type = 'CIRCLE' AND (p_center IS NULL OR p_radius <= 0) THEN
        RETURN FALSE;
    END IF;
    
    IF p_fence_type = 'POLYGON' AND p_polygon IS NULL THEN
        RETURN FALSE;
    END IF;
    
    RETURN TRUE;
END;
/

8.2 定时清理任务

-- 创建位置数据清理作业
CREATE OR REPLACE PROCEDURE cleanup_old_locations(
    p_retention_days INT DEFAULT 90
)
AS
BEGIN
    -- 删除过期的用户位置记录
    DELETE FROM user_locations 
    WHERE record_time < SYSDATE - p_retention_days;
    
    -- 归档围栏触发记录
    INSERT INTO fence_triggers_archive
    SELECT * FROM fence_triggers 
    WHERE trigger_time < SYSDATE - 365;
    
    DELETE FROM fence_triggers 
    WHERE trigger_time < SYSDATE - 365;
    
    COMMIT;
    
    -- 更新统计信息
    DBMS_STATS.GATHER_TABLE_STATS('LBS_USER', 'USER_LOCATIONS');
END;
/

-- 创建定时任务
DBMS_SCHEDULER.CREATE_JOB(
    job_name        => 'CLEANUP_LOCATIONS_JOB',
    job_type        => 'STORED_PROCEDURE',
    job_action      => 'LBS_USER.CLEANUP_OLD_LOCATIONS',
    start_date      => SYSTIMESTAMP,
    repeat_interval => 'FREQ=DAILY;BYHOUR=2',
    enabled         => TRUE
);

---

九、监控与维护

空间数据统计

-- 查看空间数据分布
SELECT 
    COUNT(*) as total_stores,
    MIN(ST_X(location)) as min_lng,
    MAX(ST_X(location)) as max_lng,
    MIN(ST_Y(location)) as min_lat,
    MAX(ST_Y(location)) as max_lat
FROM store_locations;

-- 空间索引使用统计
SELECT 
    index_name,
    num_rows,
    leaf_blocks,
    distinct_keys
FROM user_indexes 
WHERE index_type LIKE '%SPATIAL%';

-- 查询性能监控
SELECT 
    sql_text,
    executions,
    elapsed_time / 1000000 as elapsed_seconds,
    buffer_gets
FROM v$sqlarea 
WHERE sql_text LIKE '%ST_%'
ORDER BY elapsed_time DESC
LIMIT 10;

---

十、常见问题处理

10.1 坐标系转换

-- WGS84转GCJ-02（中国坐标系）函数
CREATE OR REPLACE FUNCTION wgs84_to_gcj02(
    wgs_lng FLOAT,
    wgs_lat FLOAT
) 
RETURN VARCHAR
AS
    -- 实现坐标系转换算法（这里需要实现具体转换逻辑）
    gcj_lng FLOAT;
    gcj_lat FLOAT;
BEGIN
    -- 这里添加坐标系转换算法
    -- ...
    
    RETURN CONCAT(gcj_lng, ',', gcj_lat);
END;
/

-- 创建转换视图
CREATE VIEW store_locations_gcj02 AS
SELECT 
    store_id,
    store_name,
    wgs84_to_gcj02(ST_X(location), ST_Y(location)) as gcj_coord
FROM store_locations;

10.2 性能问题排查

-- 1. 检查空间索引有效性
SELECT 
    table_name,
    index_name,
    status
FROM user_indexes 
WHERE index_type LIKE '%SPATIAL%';

-- 2. 分析空间查询执行计划
EXPLAIN PLAN FOR
SELECT * FROM store_locations 
WHERE ST_DWithin(location, ST_Point(116.407526, 39.904030), 0.01);

SELECT * FROM TABLE(DBMS_XPLAN.DISPLAY());

-- 3. 监控空间函数性能
SELECT 
    function_name,
    total_time,
    calls
FROM v$function
WHERE function_name LIKE 'ST_%';

---

总结

本教程涵盖了达梦数据库在LBS应用中的核心功能实现，包括：

1. 空间数据存储：使用达梦的空间数据类型存储地理位置信息

2. 空间查询：实现距离计算、范围查询、空间关系判断

3. 地理围栏：实现电子围栏的触发和监控

4. 性能优化：空间索引、分区表、查询优化技巧

5. 应用集成：Java和Spring Boot集成示例

6. 运维监控：数据清理、性能监控、问题排查

达梦数据库提供了完整的空间数据处理能力，可以满足大多数LBS应用的需求。在实际应用中，还需要结合具体业务场景进行优化和扩展。