mobilitydb

MobilityDB地理空间投影数据管理分析平台

概览

扩展包名版本分类许可证语言
mobilitydb1.3.0GISGPL-3.0C
ID扩展名BinLibLoadCreateTrustReloc模式
1650mobilitydb-
1651mobilitydb_datagen-
相关扩展postgis pgrouting h3_postgis timescaledb postgis_topology postgis_raster postgis_sfcgal postgis_tiger_geocoder address_standardizer
下游依赖mobilitydb_datagen

need another schema

版本

类型仓库版本PG 大版本包名依赖
EXTPGDG1.3.01817161514mobilitydbpostgis
DEBPGDG1.3.01817161514postgresql-$v-mobilitydb-
OS / PGPG18PG17PG16PG15PG14
el8.x86_64PGDG MISSPGDG MISSPGDG MISSPGDG MISSPGDG MISS
el8.aarch64PGDG MISSPGDG MISSPGDG MISSPGDG MISSPGDG MISS
el9.x86_64PGDG MISSPGDG MISSPGDG MISSPGDG MISSPGDG MISS
el9.aarch64PGDG MISSPGDG MISSPGDG MISSPGDG MISSPGDG MISS
el10.x86_64PGDG MISSPGDG MISSPGDG MISSPGDG MISSPGDG MISS
el10.aarch64PGDG MISSPGDG MISSPGDG MISSPGDG MISSPGDG MISS
d12.x86_64
d12.aarch64
d13.x86_64
d13.aarch64
u22.x86_64PGDG MISS
u22.aarch64PGDG MISS
PGDG 1.2.0
PGDG 1.2.0
PGDG 1.2.0
PGDG 1.2.0
u24.x86_64
u24.aarch64
u26.x86_64
u26.aarch64

安装

您可以直接安装 mobilitydb 扩展包的预置二进制包,首先确保 PGDG 仓库已经添加并启用:

pig repo add pgdg -u          # 添加 PGDG 仓库并更新缓存

使用 pig 或者是 apt/yum/dnf 安装扩展:

pig install mobilitydb;          # 当前活跃 PG 版本安装
pig ext install -y mobilitydb -v 18  # PG 18
pig ext install -y mobilitydb -v 17  # PG 17
pig ext install -y mobilitydb -v 16  # PG 16
pig ext install -y mobilitydb -v 15  # PG 15
pig ext install -y mobilitydb -v 14  # PG 14
apt install -y postgresql-18-mobilitydb   # PG 18
apt install -y postgresql-17-mobilitydb   # PG 17
apt install -y postgresql-16-mobilitydb   # PG 16
apt install -y postgresql-15-mobilitydb   # PG 15
apt install -y postgresql-14-mobilitydb   # PG 14

创建扩展

CREATE EXTENSION mobilitydb CASCADE;  -- 依赖: postgis

用法

来源:repo README, MobilityDB 1.3 manual, v1.3.0 release

MobilityDB 使用 temporal 和 spatio-temporal 数据类型扩展 PostgreSQL 与 PostGIS,使车辆轨迹、传感器读数和随时间变化属性等 moving object data 能够高效存储、索引和查询。

关键文档:

入门

MobilityDB 要求 PostGIS。启用两个扩展:

CREATE EXTENSION postgis;
CREATE EXTENSION mobilitydb;

Temporal 类型

MobilityDB 为基础类型提供 temporal variants:

Temporal TypeBase TypeDescription
tboolbooleanTime-varying boolean
tintintegerTime-varying integer
tfloatfloatTime-varying float
ttexttextTime-varying text
tgeometrygeometryTime-varying arbitrary geometry
tgeographygeographyTime-varying arbitrary geography
tgeompointgeometry(Point)Time-varying geometric point
tgeogpointgeography(Point)Time-varying geographic point
tnpointnetwork pointTime-varying network point
tcbuffercircular bufferTime-varying circular buffer
tposeposeTime-varying point position and orientation
trgeometryrigid geometryTime-varying rigid geometry

MobilityDB 1.3 增加了 tgeometrytgeographytcbuffertposetrgeometrytgeometrytgeography 支持 discrete 或 step interpolation,不支持任意几何的 linear interpolation。1.3 release notes 将 tcbuffertposetrgeometry 标记为 experimental。

Temporal 子类型

每种 temporal type 可按值如何随时间变化表示为不同子类型:

SubtypeDescriptionExample
Instant单个时间戳上的单个值'25.5@2025-01-01 08:00'
Sequence一个时间区间上的连续值'[25.5@08:00, 28.1@09:00, 30.0@10:00]'
SequenceSet一组互不重叠的 sequences'{[25.5@08:00, 28.1@09:00], [30.0@11:00, 31.2@12:00]}'

Sequences 使用方括号表示包含 [ 或排除 ( 边界,就像 PostgreSQL range types 一样。

创建 Temporal 值

Instant values:

SELECT tfloat '25.5@2025-06-01 08:00:00+00';
SELECT tgeompoint 'SRID=4326;Point(2.3522 48.8566)@2025-06-01 08:00:00+00';

Sequence values(continuous interpolation):

SELECT tfloat '[20.0@2025-06-01 08:00, 25.5@2025-06-01 09:00, 22.0@2025-06-01 10:00]';

Discrete sequences(stepwise interpolation):

SELECT tint 'Interp=Step;[10@2025-06-01 08:00, 20@2025-06-01 09:00, 15@2025-06-01 10:00]';

SequenceSet values:

SELECT tfloat '{[20.0@08:00, 25.5@09:00], [22.0@11:00, 28.0@12:00]}';

从组件构造:

SELECT tgeompoint_inst(ST_Point(2.3522, 48.8566, 4326), '2025-06-01 08:00+00');
SELECT tgeompoint_seq(ARRAY[
    tgeompoint_inst(ST_Point(2.3522, 48.8566, 4326), '2025-06-01 08:00+00'),
    tgeompoint_inst(ST_Point(2.2945, 48.8584, 4326), '2025-06-01 08:30+00'),
    tgeompoint_inst(ST_Point(2.3364, 48.8606, 4326), '2025-06-01 09:00+00')
]);

Temporal 操作

提取指定时间的值:

SELECT valueAtTimestamp(temp, '2025-06-01 08:30:00+00')
FROM (SELECT tfloat '[20.0@08:00, 30.0@09:00]' AS temp) t;
-- Returns 25.0 (linear interpolation)

限制到某个时间段:

SELECT atTime(trip, tstzspan '[2025-06-01 08:00, 2025-06-01 09:00]')
FROM trips;

获取 temporal value 的时间跨度:

SELECT duration(trip), startTimestamp(trip), endTimestamp(trip)
FROM trips;

Temporal comparisons:

-- Time periods when temperature exceeded 30 degrees
SELECT atValue(temperature, true)
FROM (SELECT tfloat '[20@08:00, 35@09:00, 25@10:00]' #> 30.0 AS temperature) t;

Spatial-Temporal 操作

Trajectory:提取空间路径为 geometry:

SELECT ST_AsText(trajectory(trip))
FROM trips
WHERE vehicle_id = 42;

速度计算:

-- Speed in units per second (m/s for geographic points)
SELECT speed(trip)
FROM trips
WHERE vehicle_id = 42;

轨迹长度:

SELECT length(trip)
FROM trips
WHERE vehicle_id = 42;

Space-time bounding box(stbox):

-- Get the space-time bounding box
SELECT stbox(trip)
FROM trips;

-- Construct an stbox for querying
SELECT stbox(
    ST_MakeEnvelope(2.2, 48.8, 2.4, 48.9, 4326),
    tstzspan '[2025-06-01, 2025-06-02]'
);

空间限制:区域内的值:

-- Portions of a trip within a polygon
SELECT atGeometry(trip, ST_Buffer(ST_Point(2.35, 48.86, 4326), 0.01))
FROM trips;

两个 temporal points 之间的距离:

SELECT distance(t1.trip, t2.trip)
FROM trips t1, trips t2
WHERE t1.vehicle_id = 1 AND t2.vehicle_id = 2;

最近接近距离和时间:

SELECT nearestApproachDistance(t1.trip, t2.trip),
       nearestApproachInstant(t1.trip, t2.trip)
FROM trips t1, trips t2
WHERE t1.vehicle_id = 1 AND t2.vehicle_id = 2;

索引

MobilityDB 支持 GiST 和 SP-GiST indexes,用于高效 temporal 与 spatio-temporal 查询。

Temporal types(time dimension)的 SP-GiST index:

CREATE INDEX ON measurements USING spgist(temperature);

Spatio-temporal types(space + time)的 GiST index:

CREATE INDEX ON trips USING gist(trip);

这些索引可加速 bounding box 查询、temporal overlap 检查和 spatial-temporal intersection:

-- Uses GiST index for space-time filtering
SELECT vehicle_id
FROM trips
WHERE trip && stbox(
    ST_MakeEnvelope(2.2, 48.8, 2.4, 48.9, 4326),
    tstzspan '[2025-06-01, 2025-06-02]'
);

示例:车辆跟踪

存储和查询车辆 GPS 轨迹的完整示例:

CREATE TABLE vehicles (
    vehicle_id  INT PRIMARY KEY,
    plate       TEXT,
    type        TEXT
);

CREATE TABLE trips (
    trip_id     BIGSERIAL PRIMARY KEY,
    vehicle_id  INT REFERENCES vehicles(vehicle_id),
    trip        tgeompoint,
    trip_date   DATE
);

CREATE INDEX ON trips USING gist(trip);

-- Insert a trip as a sequence of GPS points
INSERT INTO trips (vehicle_id, trip, trip_date) VALUES (
    1,
    tgeompoint_seq(ARRAY[
        tgeompoint_inst(ST_Point(2.3522, 48.8566, 4326), '2025-06-01 08:00+00'),
        tgeompoint_inst(ST_Point(2.2945, 48.8584, 4326), '2025-06-01 08:15+00'),
        tgeompoint_inst(ST_Point(2.3364, 48.8606, 4326), '2025-06-01 08:30+00'),
        tgeompoint_inst(ST_Point(2.3488, 48.8534, 4326), '2025-06-01 08:45+00')
    ]),
    '2025-06-01'
);

-- Where was vehicle 1 at 08:20?
SELECT valueAtTimestamp(trip, '2025-06-01 08:20+00')
FROM trips WHERE vehicle_id = 1 AND trip_date = '2025-06-01';

-- What was the average speed?
SELECT twAvg(speed(trip))
FROM trips WHERE vehicle_id = 1 AND trip_date = '2025-06-01';

-- Total distance traveled
SELECT length(trip)
FROM trips WHERE vehicle_id = 1 AND trip_date = '2025-06-01';

-- Get the full trajectory as a LineString
SELECT ST_AsGeoJSON(trajectory(trip))
FROM trips WHERE vehicle_id = 1 AND trip_date = '2025-06-01';

示例:Spatio-Temporal Intersection 查询

查找在给定时间窗口内经过特定区域的所有 trips:

-- Define area of interest: a circle around the Eiffel Tower
WITH area AS (
    SELECT ST_Buffer(ST_Point(2.2945, 48.8584, 4326)::geography, 500)::geometry AS geom
)
SELECT t.vehicle_id,
       t.trip_date,
       atGeometry(t.trip, a.geom) AS trip_in_area,
       length(atGeometry(t.trip, a.geom)) AS distance_in_area
FROM trips t, area a
WHERE t.trip && stbox(
    a.geom,
    tstzspan '[2025-06-01 07:00+00, 2025-06-01 10:00+00]'
)
  AND eIntersects(t.trip, a.geom)
ORDER BY t.trip_date;

聚合函数

MobilityDB 提供 temporal aggregates:

-- Time-weighted average of a temporal float
SELECT twAvg(temperature) FROM sensor_data WHERE sensor_id = 1;

-- Merge multiple temporal points into one
SELECT tUnion(trip) FROM trips WHERE vehicle_id = 1 AND trip_date = '2025-06-01';

-- Centroid of a set of temporal points at each timestamp
SELECT tCentroid(trip) FROM trips WHERE trip_date = '2025-06-01';

注意事项

  • catalog package 和 extension 都是 mobilitydb,版本 1.3.0;打包矩阵面向 PostgreSQL 14 到 18,并要求 postgis
  • v1.3.0 release 增加 PostgreSQL 18 和 PostGIS 3.6 支持,但 migration note 说明二进制格式相较 MobilityDB 1.2 已改变,因此从 1.2 升级需要 backup and restore。
  • 上游源码构建说明展示了在加载 MobilityDB 前设置 shared_preload_libraries = 'postgis-3'max_locks_per_transaction = 128。对于未使用打包默认值的集群,请验证这些设置。
  • 本地 package metadata 仍带有 curation comment need another schema;上游文档未确认必须使用单独 schema,因此在该备注解决前应避免给出 schema-specific guidance。

最后修改 2026-05-18: routine extension update (ac43610)