Build multi-tenant RAG with Neon's database-per-user model — no nosy neighbors, max isolation, minimal costs
15.2kLog InSign Up
Docs/Neon Postgres guides/Functions/JSON functions/jsonb_extract_path

Postgres jsonb_extract_path() function

Extracts a JSONB sub-object at the specified path

You can use the jsonb_extract_path function to extract the value at a specified path within a JSONB document. This approach is more performant compared to querying the entire JSONB payload and processing it on the application side. It is particularly useful when dealing with nested JSONB structures.

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

jsonb_extract_path(from_json JSONB, VARIADIC path_elems TEXT[]) -> JSONB

Example usage

To illustrate the jsonb_extract_path function in Postgres, let's consider a scenario where we have a table storing information about books. Each book has a JSONB column containing details such as title, author, and publication year. You can create the book table using the SQL statements shown below.

books

CREATE TABLE books (
    id INT,
    info JSONB
);

INSERT INTO books (id, info)
VALUES
    (1, '{"title": "The Catcher in the Rye", "author": "J.D. Salinger", "year": 1951}'),
    (2, '{"title": "To Kill a Mockingbird", "author": "Harper Lee", "year": 1960}'),
    (3, '{"title": "1984", "author": "George Orwell", "year": 1949}');
| id | info                                                                         |
|----|------------------------------------------------------------------------------|
| 1  | {"title": "The Catcher in the Rye", "author": "J.D. Salinger", "year": 1951} |
| 2  | {"title": "To Kill a Mockingbird", "author": "Harper Lee", "year": 1960}     |
| 3  | {"title": "1984", "author": "George Orwell", "year": 1949}                   |

Now, let's use the jsonb_extract_path function to extract the title and author of each book:

SELECT
    id,
    jsonb_extract_path(info, 'title') as title,
    jsonb_extract_path(info, 'author') as author
FROM books;

This query returns the following values:

| id | title                    | author           |
|----|--------------------------|------------------|
| 1  | "The Catcher in the Rye" | "J.D. Salinger"  |
| 2  | "To Kill a Mockingbird"  | "Harper Lee"     |
| 3  | "1984"                   | "George Orwell"  |

Advanced examples

Consider a products table that stores information about the products in an e-commerce system. The table schema and data are outlined below.

products

CREATE TABLE products (
    id INT,
    attributes JSONB
);

INSERT INTO products (id, attributes)
VALUES
    (1, '{"name": "Laptop", "specs": {"brand": "Dell", "RAM": "16GB", "storage": {"type": "SSD", "capacity": "512GB"}}, "tags": ["pc"]}'),
    (2, '{"name": "Smartphone", "specs": {"brand": "Google", "RAM": "8GB", "storage": {"type": "UFS", "capacity": "256GB"}}, "tags": ["android",
    "pixel"]}'),
    (3, '{"name": "Smartphone", "specs": {"brand": "Apple", "RAM": "8GB", "storage": {"type": "UFS", "capacity": "128GB"}}, "tags": ["ios", "iphone"]}');
| id     | attributes                                                                                                                                        |
|--------|---------------------------------------------------------------------------------------------------------------------------------------------------|
| 1      | {"name": "Laptop", "specs": {"brand": "Dell", "RAM": "16GB", "storage": {"type": "SSD", "capacity": "512GB"}}, "tags": ["pc"]}                    |
| 2      | {"name": "Smartphone", "specs": {"brand": "Google", "RAM": "8GB", "storage": {"type": "UFS", "capacity": "256GB"}}, "tags": ["android", "pixel"]} |
| 3      | {"name": "Smartphone", "specs": {"brand": "Apple", "RAM": "8GB", "storage": {"type": "UFS", "capacity": "128GB"}}, "tags": ["ios", "iphone"]}     |

Extract value from nested JSONB object with jsonb_extract_path

Let's use jsonb_extract_path to retrieve information about the storage type and capacity for each product, demonstrating how to extract values from a nested JSONB object.

SELECT
    id,
    jsonb_extract_path(attributes, 'specs', 'storage', 'type') as storage_type,
    jsonb_extract_path(attributes, 'specs', 'storage', 'capacity') as storage_capacity
FROM products;

This query returns the following values:

| id | storage_type | storage_capacity |
|----|--------------|------------------|
| 1  | "SSD"        | "512GB"          |
| 2  | "UFS"        | "256GB"          |
| 3  | "UFS"        | "128GB"          |

Extract values from JSON array with jsonb_extract_path

Now, let's use jsonb_extract_path to extract information about the associated tags as well, demonstrating how to extract values from a JSONB array.

SELECT
    id,
    jsonb_extract_path(attributes, 'specs', 'storage', 'type') as storage_type,
    jsonb_extract_path(attributes, 'specs', 'storage', 'capacity') as storage_capacity,
    jsonb_extract_path(attributes, 'tags', '0') as first_tag,
    jsonb_extract_path(attributes, 'tags', '1') as second_tag
FROM products;

This query returns the following values:

| id | storage_type | storage_capacity | first_tag | second_tag |
|----|--------------|------------------|-----------|------------|
| 1  | "SSD"        | "512GB"          | "pc"      |  null      |
| 2  | "UFS"        | "256GB"          | "android" | "pixel"    |
| 3  | "UFS"        | "128GB"          | "ios"     | "iphone"   |

Joining data with values extracted using jsonb_extract_path

Let's say you have two tables, employees and departments, and the employees table has a JSONB column named details that contains information about each employee's department. You want to join these tables based on the department information stored in the JSONB column.

The table schemas and data used in this example are shown below.

departments

CREATE TABLE departments (
    department_id SERIAL PRIMARY KEY,
    department_name VARCHAR(255)
);

INSERT INTO departments (department_name)
VALUES
    ('IT'),
    ('HR'),
    ('Marketing');
| department_id | department_name  |
|---------------|------------------|
|             1 | IT               |
|             2 | HR               |
|             3 | Marketing        |

employees

CREATE TABLE employees (
    employee_id SERIAL PRIMARY KEY,
    employee_name VARCHAR(255),
    details JSONB
);

INSERT INTO employees (employee_name, details)
VALUES
    ('John Doe', '{"department": "IT"}'),
    ('Jane Smith', '{"department": "HR"}'),
    ('Bob Johnson', '{"department": "Marketing"}');
| employee_id | employee_name |           details           |
|-------------|---------------|-----------------------------|
|           1 | John Doe      | {"department": "IT"}        |
|           2 | Jane Smith    | {"department": "HR"}        |
|           3 | Bob Johnson   | {"department": "Marketing"} |

You can use JOIN with jsonb_extract_path to retrieve the value to join on:

SELECT
    employees.employee_name,
    departments.department_name
FROM
    employees
JOIN
    departments ON TRIM(BOTH '"' FROM jsonb_extract_path(employees.details, 'department')::TEXT) = departments.department_name;

This query returns the following values:

| employee_name | department_name  |
|---------------|------------------|
| John Doe      | IT               |
| Jane Smith    | HR               |
| Bob Johnson   | Marketing        |

The jsonb_extract_path function extracts the value of the department key from the JSONB column in the employees table. The JOIN is then performed based on matching department names.

Handling invalid path inputs to jsonb_extract_path

jsonb_extract_path handles an invalid path by returning NULL, as in the following example:

SELECT
    id,
    jsonb_extract_path(attributes, 'speks') as storage_type
FROM products;

The query above, which specifies an invalid path ('speks' instead of 'specs'), returns NULL as shown:

id | storage_type
----+--------------
  1 |
  2 |
  3 |

Additional considerations

Performance and Indexing

The jsonb_extract_path function performs well when extracting data from JSONB documents, especially compared to extracting data in application code. It allows performing the extraction directly in the database, avoiding transferring entire JSONB documents to the application.

Indexing JSONB documents can also significantly improve jsonb_extract_path query performance when filtering data based on values extracted from JSON.

Alternative functions

  • jsonb_extract_path_text - The regular jsonb_extract_path function returns the extracted value as a JSONB object or array, preserving its JSON structure, whereas the alternative jsonb_extract_path_text function returns the extracted value as a plain text string, casting any JSONB objects or arrays to their string representations.

    Use the regular jsonb_extract_path function when you need to apply JSONB-specific functions or operators to the extracted value, requiring JSONB data types. The alternative jsonb_extract_path_text function is preferable if you need to work directly with the extracted value as a string, for text processing, concatenation, or comparison.

  • json_extract_path - The jsonb_extract_path function works with the JSONB data type, which offers a binary representation of JSON data, whereas json_extract_path takes a JSON value as an input and returns JSON too. The JSONB variant is typically more performant at query time, which is even more pronounced with larger JSON data payloads and frequent path extractions.

Resources

Need help?

Join our Discord Server to ask questions or see what others are doing with Neon. Users on paid plans can open a support ticket from the console. For more details, see Getting Support.

Previousjson_extract_path
Nextjson_extract_path_text

Last updated on

Was this page helpful?