从多列创建 JSON 个对象的数组

Question

我这里有以下数据集

id	key	category	avg_time_1	avg_time_2
1	1	10	10	20
2	1	20	30	40
3	2	10	10	50
4	2	20	60	70

Sqlfiddle.

我想创建一个查询，结果如下：

key	avg_time_1	avg_time_1
1	[{ "category": 10, "avg_time": 10},{ "category": 20, "avg_time": 20 }]	[{"category": 10, "avg_time": 20}, {"category": 20, "avg_time": 40}]
2	[{ "category": 10, "avg_time": 10},{ "category": 20, "avg_time": 60}]	[{"category": 10, "avg_time": 50}, {"category": 20, "avg_time": 70}]

这个想法只是以不同的方式将值 cols avg_time_1 和 avg_time_2 表示为 jsonb 值。

Answer 1

使用jsonb_build_object()构建素数对象并jsonb_agg()将它们聚合成json数组：

select 
    key, 
    jsonb_agg(jsonb_build_object('category', category, 'avg_time', avg_time_1)) as avg_time_1,
    jsonb_agg(jsonb_build_object('category', category, 'avg_time', avg_time_2)) as avg_time_2
from data_to_agg_json
group by key

 key |                              avg_time_1                              |                              avg_time_2                              
-----+----------------------------------------------------------------------+----------------------------------------------------------------------
   1 | [{"avg_time": 10, "category": 10}, {"avg_time": 30, "category": 20}] | [{"avg_time": 20, "category": 10}, {"avg_time": 40, "category": 20}]
   2 | [{"avg_time": 10, "category": 10}, {"avg_time": 60, "category": 20}] | [{"avg_time": 50, "category": 10}, {"avg_time": 70, "category": 20}]
(2 rows)    

Db<>Fiddle.