Question

POSTGRES查询遇到一些麻烦。

我有一列jsonb，名称为“ venture_menu”，其中包含一系列对象，如下所示：

[
    { "menu_id":"0", "menu_name":"name 1"},
    { "menu_id":"1", "menu_name":"name 2"},
    { "menu_id":"2", "menu_name":"name 3"}
]

我想进行更新，例如在由client_id列（WHERE子句）选择的特定行的menu_id为2的对象上。到目前为止，我有以下查询

UPDATE client SET venue_menu = jsonb_set(venue_menu, '{}', { "menu_id":"2", "menu_name":"name updated"}) WHERE client_id = "1";

我似乎无法弄清楚在哪里进行查询以指定要更新的对象的键名，有什么想法吗？

谢谢。

Answer 1

使用jsonb_set

update "t" set "col"= jsonb_set("cols":: JSONB,('{'|| elem_index || ',"' ||'menu_name"}')::text[],'"name updated"'::jsonb,false)
from (select pos- 1 as elem_index,"col" as "cols","client_id" as "colId"
from  "t",jsonb_array_elements("t"."col" :: JSONB) with ordinality arr(elem, pos) where elem->>'menu_id' = '2') as "tble"     
where  client_id =1

DEMO

Answer 2

您可以使用：

+----+---------------------+--------+---------+----------------+-------------+----------+---------------------------- | | date | city | weather | prior_the_same | prior_types |expected | why? +----+---------------------+--------+---------+----------------+-------------+----------+---------------------------- | 1 | 11.08.2018 00:00:00 | Ankara | Sun | 5 | 2 | 11 | 11t day since 1st time Sun | 2 | 10.08.2018 00:00:00 | Ankara | Sun | 4 | 3 | 10 | 10t day since 1st time Sun | 3 | 09.08.2018 00:00:00 | Ankara | Sun | 7 | 3 | 9 | 9th day since 1st time Sun | 4 | 08.08.2018 00:00:00 | Ankara | Sun | 7 | 3 | 8 | 8th day since 1st time Sun | 5 | 07.08.2018 00:00:00 | Ankara | Sun | 6 | 3 | 7 | 7th day since 1st time Sun | 6 | 06.08.2018 00:00:00 | Ankara | Sun | 5 | 3 | 6 | 6th day since 1st time Sun ( <5 days gap since last Sun keeps counting ) | 7 | 05.08.2018 00:00:00 | Ankara | Rain | 2 | 3 | 3 | 3rd day since 1st time Rain | 8 | 04.08.2018 00:00:00 | Ankara | Clouds | 0 | 3 | 1 | 1st day Clouds | 9 | 03.08.2018 00:00:00 | Ankara | Rain | 0 | 2 | 1 | 1st day Rain | 10 | 02.08.2018 00:00:00 | Ankara | Sun | 1 | 0 | 2 | 2nd day since 1st time Sun | 11 | 01.08.2018 00:00:00 | Ankara | Sun | 0 | 0 | 1 | 1st day Sun | 12 | 11.08.2018 00:00:00 | Cairo | Clouds | 6 | 6 | 1 | 1st time Clouds ( >5 days gap since last Clouds resets the count ) | 13 | 10.08.2018 00:00:00 | Cairo | Sun | 7 | 1 | 10 | 10t day since 1st time Sun | 14 | 09.08.2018 00:00:00 | Cairo | Sun | 7 | 1 | 9 | 9th day since 1st time Sun | 15 | 08.08.2018 00:00:00 | Cairo | Sun | 7 | 1 | 8 | 8th day since 1st time Sun | 16 | 07.08.2018 00:00:00 | Cairo | Sun | 6 | 1 | 7 | 7th day since 1st time Sun | 17 | 06.08.2018 00:00:00 | Cairo | Sun | 5 | 1 | 6 | 6th day since 1st time Sun ( <5 days gap since last Sun keeps counting ) | 18 | 05.08.2018 00:00:00 | Cairo | Clouds | 0 | 4 | 1 | 1st time Clouds | 19 | 04.08.2018 00:00:00 | Cairo | Sun | 3 | 0 | 4 | 4th day since 1st time Sun | 20 | 03.08.2018 00:00:00 | Cairo | Sun | 2 | 0 | 3 | 3rd day since 1st time Sun | 21 | 02.08.2018 00:00:00 | Cairo | Sun | 1 | 0 | 2 | 2nd day since 1st time Sun | 22 | 01.08.2018 00:00:00 | Cairo | Sun | 0 | 0 | 1 | 1st day Sun +----+---------------------+--------+---------+----------------+-------------+----------+----------------------------将数组拆分为行
jsonb_to_recordset使用拆分的行构建每个元素
jsonb_build_object将每个元素连接到一个新数组

这里是示例：

jsonb_agg

您可以在https://dbfiddle.uk/?rdbms=postgres_10&fiddle=147c0de2abd5d0df786ed636793f1cc8

上看到正在运行的示例。

Postgres：更新数组中的特定json对象

2 个答案: