PostgreSQL jsonb_set() Function

Summary: in this tutorial, you will learn how to use the PostgreSQL jsonb_set() function to replace an existing value specified by a path with a new value in a JSON document.

Introduction to the PostgreSQL jsonb_set() function

The jsonb_set() function allows you to replace an existing value specified by a path with a new value in a JSON document of the JSONB type.

More specifically, the jsonb_set() function allows you to replace an array element or key/value in a JSON object, or nested combinations of them.

Here’s the syntax of the jsonb_set() function:

jsonb_set(
   target jsonb,
   path text[],
   new_value jsonb
   [, create_missing boolean]
)

In this syntax:

• target: This is the original JSON document of the JSONB type that you want to modify.
• path: This is an array of text elements that specifies the path to the key where you want to insert or update the data.
• new_value is the new JSONB value that you want to set at the specified path.
• create_missing: This is an optional boolean parameter indicating whether you want to create missing keys if they do not exist. It defaults to true, meaning that the function will create a new key if you attempt to set it for a key that does not exist.

The jsonb_set() function returns the modified JSON document with the new_value set at a specified path.

PostgreSQL jsonb_set() function examples

Let’s explore some examples of using the PostgreSQL jsonb_set() function

1) Updating an existing element in a JSON array

The following example uses the jsonb_set() function to update an existing element in a JSON array:

SELECT jsonb_set('[1,2,3]', '{0}', '-1');

Output:

jsonb_set
------------
 [-1, 2, 3]
(1 row)

In this example:

• The original array is [1,2,3].
• The path {0} indicates the first element of the array.
• The number -1 is the new value.

The jsonb_set() function sets the first element of the array to -1 and returns the modified document.

To insert the number 4 after the 3rd element, you use a non-existing path to the 4th element as follows:

SELECT jsonb_set('[1,2,3]', '{4}', '4');

Output:

jsonb_set
--------------
 [1, 2, 3, 4]
(1 row)

2) Updating an element in a nested JSON array

The following example uses the jsonb_set() function to update an element in a nested array:

SELECT
  jsonb_set(
    '[1,2,[4,5],6]', '{2,0}', '3'
  );

Output:

jsonb_set
-------------------
 [1, 2, [3, 5], 6]
(1 row)

In this example:

• The original array is [1,2,[4,5],6].
• The path 0, 2 specifies the second element of the array which is the nested array [4,5], and 0 specifies the first element of the nested array.
• 3 is the new value.

Therefore the jsonb_set() function changes the number 4 as the first element of the nested array [4,5] to 3.

3) Updating data in a JSON object

The following example uses the jsonb_set() to update the value of a key in a JSON object:

SELECT
  jsonb_set('{"name": "Jane Doe"}', '{name}', '"Jane Smith"');

Output:

jsonb_set
------------------------
 {"name": "Jane Smith"}
(1 row)

In this example:

• \{“name”: “Jane Doe”\} is the original object.
• \{name\} is the path that indicates the name property (or key).
• “Jane Smith” is the new value to update.

Therefore, the jsonb_set() set the value of the name key in the JSON object to “Jane Smith”.

Note that if you attempt to set a key that does not exist, you’ll get an error, the jsonb_set will insert it. For example:

SELECT jsonb_set('{"name": "Jane Doe"}', '{age}', '25');

Output:

jsonb_set
---------------------------------
 {"age": 25, "name": "Jane Doe"}
(1 row)

But if you set the create_missing parameter to false, the function will not insert a new key/value pair:

SELECT
  jsonb_set(
    '{"name": "Jane Doe"}', '{age}',
    '25',
    false
  );

Output:

jsonb_set
----------------------
 {"name": "Jane Doe"}
(1 row)

4) Updating a value in a nested JSON object

The following example uses the jsonb_set() to modify a key/value pair in a nested JSON object:

SELECT
  jsonb_set(
    '{"name":"John Doe", "address" : { "city": "San Francisco"}}',
    '{address,city}', '"San Jose"'
  );

Output:

jsonb_set
-------------------------------------------------------
 {"name": "John Doe", "address": {"city": "San Jose"}}
(1 row)

In this example:

• {"name":"John Doe", "address" : { "city": "San Francisco"}} is the original JSON object.
• {address, city} is a path that specifies the address key whose value is an object and the city is the key of the address object that will be modified.
• "San Jose" is the value of the city key.

Therefore, the jsonb_set() function updates the city with the value San Jose in the address object of the JSON document.

5) Updating an element in an array of a nested object

The following example uses the jsonb_set() to update an element in an array of a nested object

SELECT
  jsonb_set(
    '{"name": "John", "skills" : ["PostgreSQL", "API"]}',
    '{skills,1}',
    '"Web Dev"'
  );

Output:

jsonb_set
-------------------------------------------------------
 {"name": "John", "skills": ["PostgreSQL", "Web Dev"]}
(1 row)

In this example:

• {"name": "John", "skills" : ["PostgreSQL", "API"]} is the original JSON object.
• {skills,1} is a path that specifies the skills key, which is an array, and 1 specifies the second element of the array.
• "Web Dev" is the new value to update.

The jsonb_set() function sets the second element of the skills array to "Web Dev".

6) Using the PostgreSQL jsonb_set() function with table data

We’ll show you how to use the jsonb_set() function to insert a new value into a JSON document and update it back to a table.

First, create a new table called employee_skills:

CREATE TABLE employee_skills(
    id INT PRIMARY KEY,
    data JSONB
);

Second, insert rows into the employee_skills table:

INSERT INTO employee_skills(id, data)
VALUES
   (1, '{"name": "John", "skills" : ["PostgreSQL", "API"]}'),
   (2, '{"name": "Jane", "skills" : ["SQL","Java"]}')
RETURNING *;

Output:

id |                       data
----+---------------------------------------------------
  1 | {"name": "John", "skills": ["PostgreSQL", "API"]}
  2 | {"name": "Jane", "skills": ["SQL", "Java"]}
(2 rows)

Third, replace the first skill in the skills array of the employee id 1 with the new skill "Web Dev":

UPDATE
  employee_skills
SET
  data = jsonb_set(
    data, '{skills,0}', '"Web Dev"'
  )
WHERE
  id = 1
RETURNING *;

Output:

id |                      data
----+------------------------------------------------
  1 | {"name": "John", "skills": ["Web Dev", "API"]}
(1 row)

Summary

• Use the jsonb_set() function to update a JSON document of the type JSONB.