Question

这是我的起点（工作查询）：

select 
    count(boot_est_pct > 0) as visited,
    count(*) as num_parcels, 
    sum((CAST(init_est_pct AS REAL) / 100) * dp.build_val) as impact,
    sum((CAST(boot_est_pct AS REAL) / 100) * dp.build_val) as pda,  
    p.st_name as street             
FROM v_damaged_parcels dp
LEFT JOIN i_parcels p ON dp.parcel = p.qpid
LEFT JOIN parcels pg ON pg.qpid = p.qpid
WHERE dp.event_id = 3
    AND (init_est_pct > 0  OR boot_est_pct > 0 OR dp.boot_type = 'IA')
    AND p.num_buildings > 0
group by p.st_name
order by p.st_name

这给了我每条街道的摘要信息。我的问题是街道可以跨越城市边界，以及团队边界，我希望能够在我的HTML页面上按城市和团队进行过滤。我想要的是一组团队，以及我的结果集中每个街道（行）的城市之一。我不想将团队和城市添加到GROUP BY。我该怎么办呢？宗地表（上面的连接之一）具有每个宗地的几何（每条街道的许多宗地）。有一个带有几何的城市表和一个带几何的团队表。

上述查询中的表格：

CREATE TABLE damaged_parcels
(
  id serial NOT NULL,
  parcel integer NOT NULL,
  st_name character(100),
  init_est_pct integer,
  boot_est_pct integer
 }

CREATE TABLE i_parcels
(
  id serial NOT NULL,
  qpid integer, -- parcel column in damaged_parcels
  st_name character(50),
  st_number character(100),
  value integer,
  num_buildings integer
 }
CREATE TABLE parcel
(
  id serial NOT NULL,
  qpid integer NOT NULL,
  gis_acres real,
  the_geom geometry,
  shapeid integer
}

团队＆amp;城市表：

CREATE TABLE event_team_region
(
  id integer NOT NULL DEFAULT nextval('event_team_region_id_seq'::regclass),
  event_id integer,
  team_id integer,
  the_geom geometry
 }

CREATE TABLE teams
(
  team_id serial NOT NULL,
  team_name character(20),
   transid integer,
  lastmodified timestamp without time zone NOT NULL DEFAULT now(),
  item_id integer NOT NULL,
  CONSTRAINT teampk PRIMARY KEY (team_id)
)

CREATE TABLE cities
(
  id integer NOT NULL DEFAULT nextval('municipal_id_seq1'::regclass),
  alt_id character(30),
  name character(60),
  code character(60),
  gis_acres real,
  the_geom geometry
 }

当前查询的结果：

+---------+---------+-----+---------+---------+
| street  | visited | num | impact  |   pda   |
+---------+---------+-----+---------+---------+
| ABACO   |       3 |  14 | 1018776 |  322707 |
|         |         |     |         |         |
| ACORN   |       1 |   1 |   23012 |  111608 |
|         |         |     |         |         |
| AMENITY |       5 |   5 |  498364 | 1145531 |
|         |         |     |         |         |
+---------+---------+-----+---------+---------+

What I want:

+---------+---------+-----+---------+---------+-----------------+--------------------+
| street  | visited | num | impact  |   pda   |     cities      |       teams        |
+---------+---------+-----+---------+---------+-----------------+--------------------+
| ABACO   |       3 |  14 | 1018776 |  322707 | Mudville,Gothem | A-team             |
|         |         |     |         |         |                 |                    |
| ACORN   |       1 |   1 |   23012 |  111608 | Orlando         | Zulu,Wonder-twerps |
|         |         |     |         |         |                 |                    |
| AMENITY |       5 |   5 |  498364 | 1145531 | Cityland        | NULL               |
|         |         |     |         |         |                 |                    |
+---------+---------+-----+---------+---------+-----------------+--------------------+

Answer 1

a_horse_with_no_name（美国！）指出了我正确的方向。他说尝试使用string_agg，但我正在使用8.4版，所以我尝试使用array_agg。这是工作查询：

    select 
        p.st_name as street,
        count(boot_est_pct > 0) as visited,
        count(*) as num_parcels, 
        sum((CAST(init_est_pct AS REAL) / 100) * dp.build_val) as impact,
        sum((CAST(boot_est_pct AS REAL) / 100) * dp.build_val) as pda,
        array_agg(trim(t.team_name)) AS team_name,
        array_agg(trim(m.name)) AS city             
    FROM v_idam_damaged_parcels dp
    LEFT JOIN idam_parcels p ON dp.parcel = p.qpid
    LEFT JOIN parcel pg on pg.qpid = p.qpid
    LEFT JOIN v_event_team_region etr on etr.event_id = dp.event_id AND st_intersects(etr.the_geom,pg.the_geom)
    LEFT JOIN v_teams t on t.team_id = etr.team_id
    LEFT JOIN municipal m ON st_intersects(m.the_geom,pg.the_geom)
    WHERE dp.event_id = 3
        AND (init_est_pct > 0  OR boot_est_pct > 0 OR dp.boot_type = 'IA')
        AND p.num_buildings > 0
    group by p.st_name, dp.boot_est_pct
    order by p.st_name

将“数组”添加到PostgreSQL查询的结果集中 - 无法想出用于表达该atm的更好方法

1 个答案: