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_populate_record

Postgres json_populate_record() function

Casts a JSON object to a record

The json_populate_record function is used to populate a record type with values from a JSON object. It is useful for parsing JSON data received from external sources, particularly when merging it into an existing record.

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_populate_record(base_record ANYELEMENT, json JSON)

This function takes two arguments: a base record of a row type (which can even be a NULL record) and a JSON object. It returns the record updated with the JSON values.

Example usage

Consider a database table that tracks employee information. When you receive employee information as JSON records, you can use json_populate_record to ingest the data into the table.

Here we create the employees table with some sample data.

CREATE TABLE employees (
    id INT,
    name TEXT,
    department TEXT,
    salary NUMERIC
);

To illustrate, we start with a NULL record and cast the input JSON payload to the employees record type.

INSERT INTO employees
SELECT *
FROM json_populate_record(
    NULL::employees,
    '{"id": "123", "name": "John Doe", "department": "Engineering", "salary": "75000"}'
)
RETURNING *;

This query returns the following result:

| id | name     | department  | salary |
|----|----------|-------------|--------|
| 123| John Doe | Engineering | 75000  |

Advanced examples

Handling partial data with json_populate_record

For data points where the JSON objects have missing keys, json_populate_record can still cast them into legible records.

Say we receive records for a bunch of employees who are known to be in Sales, but the department field is missing from the JSON payload. We can use json_populate_record with the default value specified for a field while the other fields are populated from the JSON payload, as in this example:

INSERT INTO employees
SELECT *
FROM json_populate_record(
    (1, 'ABC', 'Sales', 0)::employees,
    '{"id": "124", "name": "Jane Smith", "salary": "68000"}'
)
RETURNING *;

This query returns the following:

| id | name       | department | salary |
|----|------------|------------|--------|
| 124| Jane Smith | Sales      | 68000  |

Working with custom types in json_populate_record

The base record doesn't need to have the type of a table row and can be a custom Postgres type too. For example, here we first define a custom type address and use json_populate_record to cast a JSON object to it:

CREATE TYPE address AS (
    street TEXT,
    city TEXT,
    zip TEXT
);

SELECT *
FROM json_populate_record(
    NULL::address,
    '{"street": "123 Main St", "city": "San Francisco", "zip": "94105"}'
);

This query returns the following result:

| street     | city          | zip   |
|------------|---------------|-------|
| 123 Main St| San Francisco | 94105 |

Additional considerations

Alternative options

  • json_to_record - It can be used similarly, with a couple differences. json_populate_record can be used with a base record of a pre-defined type, whereas json_to_record needs the record type defined inline in the AS clause. Further, json_populate_record can specify default values for missing fields through the base record, whereas json_to_record must assign them NULL values.
  • json_populate_recordset - It can be used similarly to parse JSON, the difference being that it returns a set of records instead of a single record. For example, if you have an array of JSON objects, you can use json_populate_recordset to convert each object into a new row.
  • jsonb_populate_record - It has the same functionality to json_populate_record, but accepts JSONB input instead of JSON.

Resources

Last updated on

Was this page helpful?