SQL Server JSON_Modify, How to Update all?

Question

i am trying update all columns with a value with Json_Modify:

DECLARE @JSON NVARCHAR(MAX)
SET @JSON = 
N'{
"A":1,
"TMP": [    
    {"A":"VALUE1", "B": "VALUE2", "C": 1},
    {"A":"VALUE3", "B": "VALUE4", "C": 2},
    {"A":"VALUE5", "B": "VALUE6", "C": 3}]}
'

SET @JSON = JSON_MODIFY(@JSON, '$.TMP.A', 'JEJE')

SELECT * FROM OPENJSON(@JSON, '$.TMP') WITH ( A NCHAR(10), B NCHAR(10), C INT )

I need update all columns "A" with "JEJE" for example, it is not working.

Answer 1

Here are two options. Disclaimer: I am not a pro at JSON through sql server 2016, but I have hacked some stuff together.

Option 1: You are clearly creating a result set from the JSON string. Why not create the result set and then update it?

DECLARE @jsontable TABLE (A varchar(50), b varchar(50), c varchar(50))
DECLARE @JSON NVARCHAR(MAX)

SET @JSON = 
N'{
"A":1,
"TMP": [    
    {"A":"VALUE1", "B": "VALUE2", "C": 1},
    {"A":"VALUE3", "B": "VALUE4", "C": 2},
    {"A":"VALUE5", "B": "VALUE6", "C": 3}]}
'

--SET @JSON = JSON_MODIFY(@JSON, '$.TMP.A', 'JEJE')
INSERT INTO @jsontable (a,b,c)
SELECT * FROM OPENJSON(@JSON, '$.TMP') WITH ( A NCHAR(10), B NCHAR(10), C INT )

UPDATE @jsontable 
SET a = 'JEJE' 

SELECT * 
FROM @jsontable

Option 2: you can manipulate the JSON components, but you have to pass the index to the function.

DECLARE @JSON NVARCHAR(MAX)
SET @JSON = 
N'{
"A":1,
"TMP": [    
    {"A":"VALUE1", "B": "VALUE2", "C": 1},
    {"A":"VALUE3", "B": "VALUE4", "C": 2},
    {"A":"VALUE5", "B": "VALUE6", "C": 3}]}
'

SET @JSON = JSON_MODIFY(@JSON, '$.TMP[0].A', 'JEJE')
SET @JSON = JSON_MODIFY(@JSON, '$.TMP[1].A', 'JEJE')
SET @JSON = JSON_MODIFY(@JSON, '$.TMP[2].A', 'JEJE')

SELECT * FROM OPENJSON(@JSON, '$.TMP') WITH ( A NCHAR(10), B NCHAR(10), C INT )

Answer 2

I think this can be help:

CREATE PROCEDURE SP__SYSTEM___UPDATE_COLUMN_FROM_JSON
    @JSON NVARCHAR(MAX) OUTPUT,
    @COLUMN VARCHAR(50),
    @NEW_VALUE VARCHAR(50)
AS
BEGIN
    DECLARE @QUERY NVARCHAR(MAX)
    DECLARE @INDEX INT = 0

    WHILE @INDEX < (SELECT COUNT(*) FROM OPENJSON(@JSON))
    BEGIN
        SELECT @QUERY = CONCAT(N'SELECT @JSON = JSON_MODIFY(@JSON, ''$[', @INDEX, N'].',@COLUMN,''', ',@NEW_VALUE,')')
        EXEC SP_EXECUTESQL @QUERY, N'@JSON NVARCHAR(MAX) OUTPUT', @JSON = @JSON OUTPUT
        SET @INDEX = @INDEX + 1
    END
END

And you execute it like that:

BEGIN
    DECLARE @JSON VARCHAR(MAX) = '[{"ID":1,"USERID":0},{"ID":2,"USERID":0},{"ID":3,"USERID":0}]'
    EXEC SP__SYSTEM___UPDATE_COLUMN_FROM_JSON @JSON OUTPUT, 'USERID', '100'
    SELECT @JSON
END