从JSONB字段中展平聚合的键/值对?

Ric*_*ard 12 postgresql json jsonb postgresql-9.4

我在Postgres 9.4中使用下表:

     Column      ?         Type         ? Modifiers
???????????????????????????????????????????????????????????????
 id              ? integer              ? not null default
 practice_id     ? character varying(6) ? not null
 date            ? date                 ? not null
 pct_id          ? character varying(3) ?
 total_list_size ? double precision     ? not null
 star_pu         ? jsonb                ?
Run Code Online (Sandbox Code Playgroud)

我有以下查询:

SELECT date,
       AVG(total_list_size) AS total_list_size,
       json_object_agg(key, val) AS star_pu
FROM (SELECT date,
             SUM(total_list_size) AS total_list_size,
             key, SUM(value::numeric) val FROM frontend_practicelist p,
             jsonb_each_text(star_pu)
       GROUP BY date, key ) p
GROUP BY date
ORDER BY date;
Run Code Online (Sandbox Code Playgroud)

它给我一个附加到JSON对象的结果star_pu:

date            ? 2013-04-01
total_list_size ? 56025123.000000000000
star_pu         ? { "antidepressants_cost" : 180102416.8036909901975399, "antiepileptic_drugs_cost" : 296228344.171576079922216... }
Run Code Online (Sandbox Code Playgroud)

相反,我想将JSON结果展平为一系列命名空间键,因此结果如下所示:

date            ? 2013-04-01
total_list_size ? 56025123.000000000000
star_pu.antidepressants_cost        ? 180102416.8036909901975399
star_pu.antiepileptic_drugs_cost    ? 296228344.171576079922216 
...
Run Code Online (Sandbox Code Playgroud)

这可能吗?

kli*_*lin 41

这个特例

下面的函数基于表动态创建视图:

create or replace function create_totals_view(table_name text)
returns void language plpgsql as $$
declare
    s text;
begin
    execute format ($fmt$
        select string_agg(format('star_pu->>''%s'' "%s"', key, key), ',')
        from (
            select distinct key
            from %s, json_each(star_pu)
            order by 1
            ) s;
        $fmt$, '%s', '%s', table_name)
    into s;
    execute format('
        drop view if exists %s_view;
        create view %s_view as 
        select date, total_list_size, %s from %s', 
        table_name, table_name, s, table_name);
end $$;
Run Code Online (Sandbox Code Playgroud)

首先,从查询中创建一个表.

create table totals as

    SELECT date,
           AVG(total_list_size) AS total_list_size,
           json_object_agg(key, val) AS star_pu
    FROM (SELECT date,
                 SUM(total_list_size) AS total_list_size,
                 key, SUM(value::numeric) val FROM frontend_practicelist p,
                 jsonb_each_text(star_pu)
           GROUP BY date, key ) p
    GROUP BY date
    ORDER BY date;
Run Code Online (Sandbox Code Playgroud)

接下来,使用该函数,该函数将创建一个以带有_viewpostfix 的表命名的视图:

select create_totals_view('totals');
Run Code Online (Sandbox Code Playgroud)

最后,查询视图:

select * from totals_view;
Run Code Online (Sandbox Code Playgroud)

广义解(对于jsonb)

create or replace function create_jsonb_flat_view
    (table_name text, regular_columns text, json_column text)
    returns text language plpgsql as $$
declare
    cols text;
begin
    execute format ($ex$
        select string_agg(format('%2$s->>%%1$L "%%1$s"', key), ', ')
        from (
            select distinct key
            from %1$s, jsonb_each(%2$s)
            order by 1
            ) s;
        $ex$, table_name, json_column)
    into cols;
    execute format($ex$
        drop view if exists %1$s_view;
        create view %1$s_view as 
        select %2$s, %3$s from %1$s
        $ex$, table_name, regular_columns, cols);
    return cols;
end $$;
Run Code Online (Sandbox Code Playgroud)

用法:

create table example (id int, name text, params jsonb);
insert into example values
(1, 'Anna', '{"height": 175, "weight": 55}'),
(2, 'Bob', '{"age": 22, "height": 188}'),
(3, 'Cindy', '{"age": 25, "weight": 48, "pretty": true}');

select create_jsonb_flat_view('example', 'id, name', 'params');

select * from example_view;

 id | name  | age | height | pretty | weight 
----+-------+-----+--------+--------+--------
  1 | Anna  |     | 175    |        | 55
  2 | Bob   | 22  | 188    |        | 
  3 | Cindy | 25  |        | true   | 48
(3 rows)
Run Code Online (Sandbox Code Playgroud)

  • 到目前为止,这是我在堆栈溢出时见过的最有用,最深刻的答案。将jsonb字段展平到这样的视图中的一般功能应该是基于此处概述的技术的postgresql的一部分,因为它解决了jsonb字段出现的各种问题。 (3认同)
  • @adilapapaya - 一个简单的方法:`select create_jsonb_flat_view('example','id,name','params1 || params2');` (2认同)