查找 SQL 中分组行程的总距离

Question

我有一个数据显示出租车服务的个人和团体行程

|----------------------------------------------------------------------------------------------------------------------------------------------|
|  Trip_id  |   Trip_Created_Time    |   start_lat     |   start_lon    |     end_lat    |    end_lon   | trip_updated_time   |    Is_Group    |
|----------------------------------------------------------------------------------------------------------------------------------------------|
|    1      |   2021-07-01 17:29:51  |     81.91892    |  -42.19823     |   81.90281     |   -42.38918  | 2021-07-01 17:35:21 |      YES       |
|    2      |   2021-07-01 17:31:52  |     81.91892    |  -42.46920     |   81.97392     |   -42.37819  | 2021-07-01 17:52:51 |      YES       |
|    3      |   2021-07-02 21:50:51  |     81.91892    |  -42.01936     |   81.18937     |   -42.01967  | 2021-07-02 22:09:09 |      NO        |
|    4      |   2021-07-02 23:31:41  |     81.91892    |  -42.47821     |   81.01792     |   -42.17839  | 2021-07-02 23:41:51 |      NO        |
|    5      |   2021-09-21 20:12:54  |     81.91892    |  -42.47821     |   81.63829     |   -42.67292  | 2021-09-21 20:42:54 |      YES       |
|    6      |   2021-09-21 20:15:21  |     81.91892    |  -42.47821     |   81.62819     |   -42.01927  | 2021-09-21 20:59:21 |      YES       |
|    7      |   2021-09-21 20:17:23  |     81.91892    |  -42.47821     |   81.03926     |   -42.36284  | 2021-09-21 21:02:21 |      YES       |
|    8      |   2021-11-01 02:41:41  |     81.91892    |  -42.47821     |   81.36292     |   -42.47682  | 2021-07-02 23:41:51 |      NO        |
|    9      |   2021-12-21 19:19:41  |     81.91892    |  -42.47821     |   81.23671     |   -42.93628  | 2021-07-02 23:41:51 |      NO        |
|----------------------------------------------------------------------------------------------------------------------------------------------|

组 = 两个或更多用户从同一位置开始但目的地不同

我正在尝试计算分组和非分组旅行从开始 lat/lon 到结束 lat/lon 的距离

这是我的尝试：

select *,
       case when is_group = 'NO'
            then haversine(start_lat, start_lon, end_lat, end_lon)
        when is_group = 'YES'
            then NULL
       end as trip_distance
from my_table

当前输出如下：

|---------------------------------------------------------------------------------------------------------------------------------------------------------------|
|  Trip_id  |   Trip_Created_Time    |   start_lat     |   start_lon    |     end_lat    |    end_lon   | trip_updated_time   |    Is_Group    |  trip_distance |
|---------------------------------------------------------------------------------------------------------------------------------------------------------------|
|    1      |   2021-07-01 17:29:51  |     81.91892    |  -42.19823     |   81.90281     |   -42.38918  | 2021-07-01 17:35:21 |      YES       |     NULL       |
|    2      |   2021-07-01 17:31:52  |     81.91892    |  -42.46920     |   81.97392     |   -42.37819  | 2021-07-01 17:52:51 |      YES       |     NULL       |
|    3      |   2021-07-02 21:50:51  |     81.91892    |  -42.01936     |   81.18937     |   -42.01967  | 2021-07-02 22:09:09 |      NO        |    1.289247    |
|    4      |   2021-07-02 23:31:41  |     81.91892    |  -42.47821     |   81.01792     |   -42.17839  | 2021-07-02 23:41:51 |      NO        |    0.387922    |
|    5      |   2021-09-21 20:12:54  |     81.91892    |  -42.47821     |   81.63829     |   -42.67292  | 2021-09-21 20:42:54 |      YES       |     NULL       |
|    6      |   2021-09-21 20:15:21  |     81.91892    |  -42.47821     |   81.62819     |   -42.01927  | 2021-09-21 20:59:21 |      YES       |     NULL       |
|    7      |   2021-09-21 20:17:23  |     81.91892    |  -42.47821     |   81.03926     |   -42.36284  | 2021-09-21 21:02:21 |      YES       |     NULL       |
|    8      |   2021-11-01 02:41:41  |     81.91892    |  -42.47821     |   81.36292     |   -42.47682  | 2021-07-02 23:41:51 |      NO        |    3.29181     |
|    9      |   2021-12-21 19:19:41  |     81.91892    |  -42.47821     |   81.23671     |   -42.93628  | 2021-07-02 23:41:51 |      NO        |    0.29822     |
|---------------------------------------------------------------------------------------------------------------------------------------------------------------|

我如何计算 trip_distance 如果它在一个组中，即 is_group = 'YES'

编辑：分组的最终输出距离应该是该组中所有行程的距离总和，即A->B->C = A+B+C

|---------------------------------------------------------------------------------------------------------------------------------------------------------------|
|  Trip_id  |   Trip_Created_Time    |   start_lat     |   start_lon    |     end_lat    |    end_lon   | trip_updated_time   |    Is_Group    |  trip_distance |
|---------------------------------------------------------------------------------------------------------------------------------------------------------------|
|    1      |   2021-07-01 17:29:51  |     81.91892    |  -42.19823     |   81.90281     |   -42.38918  | 2021-07-01 17:35:21 |      YES       |    1.28463     | <--- 
|    2      |   2021-07-01 17:31:52  |     81.91892    |  -42.46920     |   81.97392     |   -42.37819  | 2021-07-01 17:52:51 |      YES       |    1.28463     | <--- These two total distances are same since grouped
|    3      |   2021-07-02 21:50:51  |     81.91892    |  -42.01936     |   81.18937     |   -42.01967  | 2021-07-02 22:09:09 |      NO        |    1.289247    |
|    4      |   2021-07-02 23:31:41  |     81.91892    |  -42.47821     |   81.01792     |   -42.17839  | 2021-07-02 23:41:51 |      NO        |    0.387922    |
|    5      |   2021-09-21 20:12:54  |     81.91892    |  -42.47821     |   81.63829     |   -42.67292  | 2021-09-21 20:42:54 |      YES       |    4.38921     | <---
|    6      |   2021-09-21 20:15:21  |     81.91892    |  -42.47821     |   81.62819     |   -42.01927  | 2021-09-21 20:59:21 |      YES       |    4.38921     | <---
|    7      |   2021-09-21 20:17:23  |     81.91892    |  -42.47821     |   81.03926     |   -42.36284  | 2021-09-21 21:02:21 |      YES       |    4.38921     | <--- These three total distances are same since grouped
|    8      |   2021-11-01 02:41:41  |     81.91892    |  -42.47821     |   81.36292     |   -42.47682  | 2021-07-02 23:41:51 |      NO        |    3.29181     |
|    9      |   2021-12-21 19:19:41  |     81.91892    |  -42.47821     |   81.23671     |   -42.93628  | 2021-07-02 23:41:51 |      NO        |    0.29822     |
|---------------------------------------------------------------------------------------------------------------------------------------------------------------|

Answer 1

如果我没理解错的话，你需要上一站的终点来计算到下一站的距离。我会尝试这样的事情。

select Trip_id,
       Trip_Created_Time,
       trip_updated_time,
       start_lat,
       start_lon,
       lag(end_lat, 1, 0) over (partition by start_lat, date_trunc(hour, Trip_Created_Time)  order by Trip_Created_Time, trip_updated_time) as mid_point_lat,
       lag(end_lon, 1, 0) over (partition by start_lat, date_trunc(hour, Trip_Created_Time)  order by Trip_Created_Time, trip_updated_time) as mid_point_lon,
       end_lat,
       end_lon,
       Is_Group,
       case when is_group = 'NO' or (Is_Group = 'YES' and mid_point_lat = 0)
            then haversine(start_lat, start_lon, end_lat, end_lon)
        when is_group = 'YES' and mid_point_lat != 0
            then haversine(mid_point_lat, mid_point_lon, end_lat, end_lon)
       end as trip_distance
from my_table
order by Trip_id ;

以上查询将为您提供每次旅行的真实距离。最后一步是对这些进行分组并对距离求和

Answer 2

所以你的例子 data/SQL 不会产生你的例子输出：

WITH fake_data AS (
    SELECT * FROM VALUES
        ( 1, '2021-07-01 17:29:51', 81.91892, -42.19823, 81.90281, -42.38918, '2021-07-01 17:35:21', 'YES', 1),
        ( 2, '2021-07-01 17:31:52', 81.91892, -42.46920, 81.97392, -42.37819, '2021-07-01 17:52:51', 'YES', 1),
        ( 3, '2021-07-02 21:50:51', 81.91892, -42.01936, 81.18937, -42.01967, '2021-07-02 22:09:09', 'NO', null),
        ( 4, '2021-07-02 23:31:41', 81.91892, -42.47821, 81.01792, -42.17839, '2021-07-02 23:41:51', 'NO', null),
        ( 5, '2021-09-21 20:12:54', 81.91892, -42.47821, 81.63829, -42.67292, '2021-09-21 20:42:54', 'YES', 2),
        ( 6, '2021-09-21 20:15:21', 81.91892, -42.47821, 81.62819, -42.01927, '2021-09-21 20:59:21', 'YES', 2),
        ( 7, '2021-09-21 20:17:23', 81.91892, -42.47821, 81.03926, -42.36284, '2021-09-21 21:02:21', 'YES', 2),
        ( 8, '2021-11-01 02:41:41', 81.91892, -42.47821, 81.36292, -42.47682, '2021-07-02 23:41:51', 'NO', null),
        ( 9, '2021-12-21 19:19:41', 81.91892, -42.47821, 81.23671, -42.93628, '2021-07-02 23:41:51', 'NO', null)
        t(trip_id, trip_created_time, start_lat, start_lon, end_lat, end_lon, trip_updated_time, is_group, group_id)
)
select *,
       case when is_group = 'NO'
            then haversine(start_lat, start_lon, end_lat, end_lon)
        when is_group = 'YES'
            then NULL
       end as trip_distance
from fake_data;

创建：

TRIP_ID	TRIP_CREATED_TIME	START_LAT	START_LON	END_LAT	END_LON	TRIP_UPDATED_TIME	IS_GROUP	GROUP_ID	TRIP_DISTANCE
1	2021-07-01 17:29:51	81.91892	-42.19823	81.90281	-42.38918	2021-07-01 17:35:21	YES	1
2	2021-07-01 17:31:52	81.91892	-42.4692	81.97392	-42.37819	2021-07-01 17:52:51	YES	1
3	2021-07-02 21:50:51	81.91892	-42.01936	81.18937	-42.01967	2021-07-02 22:09:09	NO		81.122258891
4	2021-07-02 23:31:41	81.91892	-42.47821	81.01792	-42.17839	2021-07-02 23:41:51	NO		100.308299209
5	2021-09-21 20:12:54	81.91892	-42.47821	81.63829	-42.67292	2021-09-21 20:42:54	YES	2
6	2021-09-21 20:15:21	81.91892	-42.47821	81.62819	-42.01927	2021-09-21 20:59:21	YES	2
7	2021-09-21 20:17:23	81.91892	-42.47821	81.03926	-42.36284	2021-09-21 21:02:21	YES	2
8	2021-11-01 02:41:41	81.91892	-42.47821	81.36292	-42.47682	2021-07-02 23:41:51	NO		61.824383293
9	2021-12-21 19:19:41	81.91892	-42.47821	81.23671	-42.93628	2021-07-02 23:41:51	NO		76.223649989

但是如果我们假设这些是有效的：

你描述的想要的（解构的）看起来像（如果提供了分组 ID）：

SELECT 
    trip_id
    ,trip_created_time
    ,start_lat
    ,start_lon
    ,end_lat
    ,end_lon 
    ,trip_updated_time
    ,is_group
    ,round(haversine(start_lat, start_lon, end_lat, end_lon),3) as dist_km
    ,sum(dist_km) over (partition by group_id) as group_sum_km
    ,iff(is_group='YES', group_sum_km, dist_km) as result
FROM fake_data
ORDER BY 1
;

给出：

TRIP_ID	TRIP_CREATED_TIME	START_LAT	START_LON	END_LAT	END_LON	TRIP_UPDATED_TIME	IS_GROUP	DIST_KM	GROUP_SUM_KM	RESULT
1	2021-07-01 17:29:51	81.91892	-42.19823	81.90281	-42.38918	2021-07-01 17:35:21	YES	3.484	9.762	9.762
2	2021-07-01 17:31:52	81.91892	-42.4692	81.97392	-42.37819	2021-07-01 17:52:51	YES	6.278	9.762	9.762
3	2021-07-02 21:50:51	81.91892	-42.01936	81.18937	-42.01967	2021-07-02 22:09:09	NO	81.122	319.478	81.122
4	2021-07-02 23:31:41	81.91892	-42.47821	81.01792	-42.17839	2021-07-02 23:41:51	NO	100.308	319.478	100.308
5	2021-09-21 20:12:54	81.91892	-42.47821	81.63829	-42.67292	2021-09-21 20:42:54	YES	31.358	162.332	162.332
6	2021-09-21 20:15:21	81.91892	-42.47821	81.62819	-42.01927	2021-09-21 20:59:21	YES	33.142	162.332	162.332
7	2021-09-21 20:17:23	81.91892	-42.47821	81.03926	-42.36284	2021-09-21 21:02:21	YES	97.832	162.332	162.332
8	2021-11-01 02:41:41	81.91892	-42.47821	81.36292	-42.47682	2021-07-02 23:41:51	NO	61.824	319.478	61.824
9	2021-12-21 19:19:41	81.91892	-42.47821	81.23671	-42.93628	2021-07-02 23:41:51	NO	76.224	319.478	76.224

可以压缩为：

SELECT 
    trip_id
    ,trip_created_time
    ,start_lat
    ,start_lon
    ,end_lat
    ,end_lon 
    ,trip_updated_time
    ,is_group
    , iff(is_group='YES'
        ,sum(haversine(start_lat, start_lon, end_lat, end_lon)) over (partition by group_id)
        ,haversine(start_lat, start_lon, end_lat, end_lon)
      ) as result
FROM fake_data
ORDER BY 1
;

查找 SQL 中分组行程的总距离

Finding the total distance of a grouped trips in SQL

sql

distance

snowflake-cloud-data-platform