JSON_OBJECTAGG (Transact-SQL)

Applies to: SQL Server 2025 (17.x) Preview Azure SQL Database Azure SQL Managed Instance SQL database in Microsoft Fabric Preview

The JSON_OBJECTAGG syntax constructs a JSON object from an aggregation of SQL data or columns. JSON_OBJECTAGG can also be used in a SELECT statement with GROUP BY GROUPING SETS clause.

The key/value pairs can be specified as input values, column, variable references.

To create a JSON array from an aggregate instead, use JSON_ARRAYAGG (Transact-SQL).

Transact-SQL syntax conventions

Syntax

JSON_OBJECTAGG ( json_key_value [ json_null_clause ] )

json_key_value ::= <json_name> : <value_expression>

json_null_clause ::= NULL ON NULL | ABSENT ON NULL

Arguments

json_key_value

The key / value pair of the JSON object.

json_null_clause

Optional. Omits the entire property of an object if the value is NULL, or use JSON null as property value. If omitted, NULL ON NULL is default.

Examples

A. Return JSON object with one key

The following example returns a JSON object with one key and null value.

select JSON_OBJECTAGG ( 'key':null )

Here's the result set.

{"key":null}

B. Construct JSON object from result set

The following example constructs a JSON object with three properties from a result set.

SELECT JSON_OBJECTAGG( c1:c2 )
FROM (
    VALUES('key1', 'c'), ('key2', 'b'), ('key3','a')
) AS t(c1, c2);

Here's the result set.

{"key1":"c","key2":"b","key3":"a"}

C. Return result with two columns

The following example returns a result with two columns. The first column contains the object_id value. The second column contains a JSON object where the key is the column name and value is the column_id.

SELECT TOP(5) c.object_id, JSON_OBJECTAGG(c.name:c.column_id) AS columns
  FROM sys.columns AS c
 GROUP BY c.object_id;

Here's the result set.

D. Return aggregated result with four columns

The following example returns a result with four columns from a SELECT statement containing SUM and JSON_OBJECTAGG aggregates with GROUP BY GROUPING SETS. The first two columns return the id and type column value. The third column total_amount returns the value of SUM aggregate on the amount column. The fourth column json_total_name_amount returns the value of JSON_OBJECTAGG aggregate on the name and amount columns.

WITH T
AS
(
SELECT *
  FROM (VALUES(1, 'k1', 'a', 2), (1, 'k2', 'b', 3), (1, 'k3', 'b', 4),
              (2, 'j1', 'd', 7) , (2, 'j2', 'd', 9)) AS b(id, name, type, amount)
)
SELECT id,
       type,
       SUM(amount) as total_amount,
       JSON_OBJECTAGG(name:amount) as json_total_name_amount
  FROM T
 GROUP BY GROUPING SETS((id), (type), (id, type), ());

Here's the result set.

Related content

• JSON Path Expressions
• JSON data in SQL Server
• JSON_ARRAYAGG (Transact-SQL)