suppose you have
create schema tmp
go
create table tmp.Properties
(
ParentId uniqueidentifier not null,
PropertyName nvarchar(20) not null,
PropertyValue nvarchar(100) null,
primary key (ParentId,PropertyName)
)
go
create table tmp.FullData
(
ParentId uniqueidentifier not null,
Properties nvarchar(max) null,
primary key (ParentId)
)
go
declare @id1 uniqueidentifier = 'F1935D6A-D5A6-4FA1-ACF4-BA3858804CEC',
@id2 uniqueidentifier = 'F1935D6B-D5A6-4FA1-ACF4-BA3858804CEC'
insert into tmp.Properties
values
(@id1, 'FirstName', 'Luke'),
(@id1, 'LastName', 'Skywalker'),
(@id2, 'FirstName', 'Han'),
(@id2, 'LastName', 'Solo')
please consider that:
how can I fill tmp.FullData with:
ParentId Properties
------------------------------------ ------------------------------------------------
F1935D6A-D5A6-4FA1-ACF4-BA3858804CEC { "FirstName": "Luke", "LastName": "Skywalker" }
F1935D6B-D5A6-4FA1-ACF4-BA3858804CEC { "FirstName": "Han", "Test1": "Solo" }
I tried
insert into tmp.FullData (ParentId, Properties)
select distinct ParentId, '{}' from tmp.Properties
update f
set Properties = json_modify(Properties, 'append $.' + p.PropertyName, p.PropertyValue)
from tmp.FullData f
cross join tmp.Properties p
but as you know/imagine
Msg 13610, Level 16, State 2, Line 39
The argument 2 of the "JSON_MODIFY" must be a string literal.
any other option? thanks in advance
答案 0 :(得分:0)
略过更新并直接插入tmp.FullData:
INSERT INTO tmp.FullData (ParentId, Properties)
SELECT
ParentId
,'{'
+ STUFF((
SELECT ',' + '"' + PropertyName + '":"' + PropertyValue + '"'
FROM tmp.Properties a
WHERE a.ParentId = p.ParentId
FOR XML PATH(''), TYPE
).value('.', 'VARCHAR(MAX)'), 1, 1, ''
) + '}' AS Properties
FROM tmp.Properties p
GROUP BY ParentId;