Postgresql and jsonb - inserting a key/value into a multi-level array - sql

Question

Very similar to this post, but I struggle to adapt from their solution..
My table : public.challenge, column lines JSONB
My initial JSON in lines :
[
{
"line": 1,
"blocs": [
{
"size": 100,
"name": "abc"
},
{
"size": 100,
"name": "def"
},
{
"size": 100,
"name": "ghi"
}
]
},
{
"line": 2,
"blocs": [
{
"size": 100,
"name": "xyz"
}
]
}
]
Desired update :
[
{
"line": 1,
"blocs": [
{
"size": 100,
"name": "abc",
"type": "regular"
},
{
"size": 100,
"name": "def",
"type": "regular"
},
{
"size": 100,
"name": "ghi",
"type": "regular"
}
]
},
{
"line": 2,
"blocs": [
{
"size": 100,
"name": "xyz",
"type": "regular"
}
]
}
]
So basically I need to add the type key+value in every object of blocs, for each element of the root array.
My unsuccessful attempt looks like this :
UPDATE public.challenge SET lines = jsonb_set(lines, '{}', (
SELECT jsonb_set(line, '{blocs}', (
SELECT jsonb_agg( bloc || '{"type":"regular"}' )
FROM jsonb_array_elements(line->'{blocs}') bloc
))
FROM jsonb_array_elements(lines) line
))
;
(currently it sets the whole column as null, maybe due to jsonb_set(lines, '{}' while my json begins as an array ?)
Thanks!

Answer 1

Use jsonb_array_elements to unnest all the array elements and then add the required json and use jsonb_agg to aggregate it again:
with cte as
(select id,
jsonb_agg(jsonb_set(val1,
'{blocs}',
(select jsonb_agg(arr2 || '{"type": "regular"}')
from jsonb_array_elements(arr1.val1 - >
'blocs') arr2)))
from challenge,
jsonb_array_elements(lines) arr1(val1)
group by 1)
update challenge
set lines = (cte.jsonb_agg)
from cte
where challenge.id = cte.id
DEMO