Build multi-tenant RAG with Neon's database-per-user model — no nosy neighbors, max isolation, minimal costs
Docs/Neon Postgres guides/Functions/JSON functions/json_object

Postgres json_object() function

Creates a JSON object from key-value pairs

The json_object function in Postgres is used to create a JSON object from a set of key-value pairs. It is particularly useful when you need to generate JSON data dynamically from existing table data or input parameters.

Try it on Neon!

Neon is Serverless Postgres built for the cloud. Explore Postgres features and functions in our user-friendly SQL editor. Sign up for a free account to get started.

Sign Up

Function signature

json_object(keys TEXT[], values TEXT[]) -> JSON
-- or --
json_object(keys_values TEXT[]) -> JSON

This function takes two text arrays as input: one for keys and one for values. Both arrays must have the same number of elements, as each key is paired with the corresponding value to construct the JSON object.

Alternatively, you can pass a single text array containing both keys and values. In this case, alternate elements in the array are treated as keys and values, respectively.

Example usage

Consider a scenario where you run a library and have a table that tracks details for each book.

The table with some sample data can be set up as shown:

-- Test database table for a bookstore inventory
CREATE TABLE book_inventory (
    book_id INT,
    title TEXT,
    author TEXT,
    price NUMERIC,
    genre TEXT
);

-- Inserting some test data into `book_inventory`
INSERT INTO book_inventory VALUES
(101, 'The Great Gatsby', 'F. Scott Fitzgerald', 18.99, 'Classic'),
(102, 'Invisible Man', 'Ralph Ellison', 15.99, 'Novel');

When querying this dataset, the frontend client might want to present the data in a different way. Say you want the catalog information just as the list of book names while combining the rest of the fields into a single metadata attribute. You can do so as shown here:

SELECT book_id, title, json_object(
  ARRAY['author', 'genre'],
  ARRAY[author, genre]
) AS metadata
FROM book_inventory;

This query returns the following result:

| book_id | title            | metadata                                   |
|---------|------------------|--------------------------------------------|
| 101     | The Great Gatsby | {"author" : "F. Scott Fitzgerald",         |
|         |                  |  "genre" : "Classic"}                      |
| 102     | Invisible Man    | {"author" : "Ralph Ellison",               |
|         |                  |  "genre" : "Novel"}                        |

Advanced examples

Creating nested JSON objects with json_object

You could use json_object to create nested JSON objects for representing more complex data. However, since json_object only expects text values for each key, we will need to combine it with other JSON functions like json_build_object. For example:

SELECT json_build_object(
  'title', title,
  'author', json_object(ARRAY['name', 'genre'], ARRAY[author, genre])
) AS book_info
FROM book_inventory;

This query returns the following result:

| book_info                                                                                        |
|--------------------------------------------------------------------------------------------------|
| {"title" : "The Great Gatsby", "author" : {"name" : "F. Scott Fitzgerald", "genre" : "Classic"}} |
| {"title" : "Invisible Man", "author" : {"name" : "Ralph Ellison", "genre" : "Novel"}}            |

Additional considerations

Gotchas and footguns

  • Ensure both keys and values arrays have the same number of elements. Mismatched arrays will result in an error. Or, if passing in a single key-value array, ensure that the array has an even number of elements.
  • Be aware of data type conversions. Since json_object expects text arrays, you may need to explicitly cast non-text data types to text.

Alternative functions

  • jsonb_object - Same functionality as json_object, but returns a JSONB object instead of JSON.
  • row_to_json - It can be used to create a JSON object from a table row (or a row of a composite type) without needing to specify keys and values explicitly. Although, it is less flexible than json_object since all fields in the row are included in the JSON object.
  • json_build_object - Similar to json_object, but allows for more flexibility in constructing the JSON object, as it can take a variable number of arguments in the form of key-value pairs.
  • json_object_agg - It is used to aggregate the key-value pairs from multiple rows into a single JSON object. In contrast, json_object outputs a JSON object for each row.

Resources

Last updated on

Was this page helpful?