JSONB modification functions in PostgreSQL

Introduction

JSONB modification functions let you change parts of JSON data stored in a database easily.

You want to update a specific value inside a JSON document without replacing the whole thing.

You need to add a new key and value to a JSON object stored in a table.

You want to remove a key from a JSON object inside your database.

You want to merge two JSON objects together inside a column.

You want to replace a nested value inside a JSONB column.

Syntax

PostgreSQL

jsonb_set(target jsonb, path text[], new_value jsonb, create_missing boolean DEFAULT true) RETURNS jsonb
jsonb_insert(target jsonb, path text[], new_value jsonb, insert_after boolean DEFAULT false) RETURNS jsonb
jsonb_strip_nulls(jsonb) RETURNS jsonb
jsonb_delete_path(target jsonb, path text[]) RETURNS jsonb

jsonb_set updates or adds a value at the specified path.

path is an array of keys or indexes to locate where to change inside the JSON.

Examples

Updates the value of age to 31.

PostgreSQL

SELECT jsonb_set('{"name": "John", "age": 30}'::jsonb, '{age}', '31'::jsonb);

Adds a new key city with value New York.

PostgreSQL

SELECT jsonb_set('{"name": "John"}'::jsonb, '{city}', '"New York"'::jsonb, true);

Inserts green before index 1 in the colors array.

PostgreSQL

SELECT jsonb_insert('{"colors": ["red", "blue"]}'::jsonb, '{colors,1}', '"green"'::jsonb, false);

Removes keys with null values.

PostgreSQL

SELECT jsonb_strip_nulls('{"a":1, "b":null}'::jsonb);

Sample Program

This query shows three JSONB modifications: updating age, adding a new key country, and removing null values.

PostgreSQL

WITH data AS (
  SELECT '{"user": {"name": "Alice", "age": 25, "city": "Boston"}}'::jsonb AS info
)
SELECT
  jsonb_set(info, '{user,age}', '26'::jsonb) AS updated_age,
  jsonb_set(info, '{user,country}', '"USA"'::jsonb, true) AS added_country,
  jsonb_strip_nulls('{"a":1, "b":null}'::jsonb) AS no_nulls
FROM data;

Important Notes

Paths in jsonb_set use curly braces and commas, like {user,age}.

Use jsonb_strip_nulls to clean JSON objects by removing keys with null values.

Remember that JSONB functions return new JSONB values; they do not change data in place unless used in an UPDATE statement.

Summary

JSONB modification functions help update, add, or remove parts of JSON data inside PostgreSQL.

They use paths to locate where to make changes inside nested JSON objects or arrays.

These functions return new JSONB values, so use them in SELECT or UPDATE queries to see changes.

Practice

1. What does the PostgreSQL function jsonb_set do when used on a JSONB column?

easy

A. It creates a new JSONB column in the table.

B. It deletes the entire JSONB column from the table.

C. It converts JSONB data into plain text format.

D. It updates or adds a value at a specified path inside the JSONB data.

Solution

1. Step 1: Understand the purpose of jsonb_set

   The function jsonb_set is designed to update or add a value at a specific location inside JSONB data using a path.

2. Step 2: Compare with other options

   The other options describe actions unrelated to jsonb_set's purpose, such as creating or deleting columns or converting data types.

3. Final Answer:

   It updates or adds a value at a specified path inside the JSONB data. -> Option D

4. Quick Check:

   jsonb_set updates JSONB data [OK]

Hint: jsonb_set changes or adds JSONB values at paths [OK]

Common Mistakes:

• Thinking jsonb_set deletes data
• Confusing jsonb_set with data type conversion
• Assuming jsonb_set creates new columns

2. Which of the following is the correct syntax to remove a key named "age" from a JSONB column data in PostgreSQL?

easy

A. UPDATE table SET data = data - 'age';

B. UPDATE table SET data = jsonb_set(data, '{age}', NULL);

C. UPDATE table SET data = data + 'age';

D. UPDATE table SET data = jsonb_remove(data, 'age');

Solution

1. Step 1: Identify the correct operator for key removal

   In PostgreSQL, the minus operator (-) removes a key from JSONB data, so data - 'age' removes the key "age".

2. Step 2: Check other options for correctness

   UPDATE table SET data = jsonb_set(data, '{age}', NULL); tries to set the key to NULL but does not remove it. UPDATE table SET data = data + 'age'; adds a key, which is incorrect. UPDATE table SET data = jsonb_remove(data, 'age'); uses a non-existent function jsonb_remove.

3. Final Answer:

   UPDATE table SET data = data - 'age'; -> Option A

4. Quick Check:

   Minus operator removes JSONB key [OK]

Hint: Use minus (-) operator to remove JSONB keys [OK]

Common Mistakes:

• Using jsonb_set with NULL to remove keys
• Trying to add keys to remove them
• Using non-existent functions like jsonb_remove

3. Given the JSONB column data with value '{"name": "Alice", "details": {"age": 30}}', what will be the result of this query?

SELECT jsonb_set(data, '{details,age}', '31', false) FROM table;

medium

A. {"name": "Alice", "details": {"age": 31}}

B. {"name": "Alice", "details": {"age": "31"}}

C. {"name": "Alice", "details": {"age": 30}}

D. {"name": "Alice", "details": {"age": null}}

Solution

1. Step 1: Understand jsonb_set parameters

   The function jsonb_set updates the value at path '{details,age}' to '31' (a JSON string) because the new value is given as text, not as a number.

2. Step 2: Check the effect of the 'create_missing' flag

   The last parameter is false, so it updates existing key without creating new ones. The age key is updated to the string '31', not the number 31.

3. Final Answer:

   {"name": "Alice", "details": {"age": "31"}} -> Option B

4. Quick Check:

   jsonb_set updates value as text unless cast [OK]

Hint: jsonb_set inserts text values unless cast to JSONB [OK]

Common Mistakes:

• Assuming '31' is numeric without casting
• Expecting no change when create_missing is false
• Confusing JSON string and number types

4. You want to update a nested key in a JSONB column but your query

UPDATE table SET data = jsonb_set(data, '{info,score}', 100);

throws an error. What is the likely cause?

medium

A. The new value 100 is not cast to JSONB, causing a type error.

B. The path '{info,score}' is invalid syntax and must be a string.

C. jsonb_set cannot update nested keys, only top-level keys.

D. The data column is not JSONB type.

Solution

1. Step 1: Check the new value type in jsonb_set

   The third argument must be JSONB type. Passing 100 as an integer without casting causes a type error.

2. Step 2: Verify other options

   The path '{info,score}' is invalid syntax and must be a string. is incorrect because the path syntax is valid. jsonb_set cannot update nested keys, only top-level keys. is false; jsonb_set supports nested keys. The data column is not JSONB type. might cause error but question implies the error is from the query syntax.

3. Final Answer:

   The new value 100 is not cast to JSONB, causing a type error. -> Option A

4. Quick Check:

   jsonb_set value must be JSONB type [OK]

Hint: Cast new value to JSONB to avoid type errors [OK]

Common Mistakes:

• Passing plain integers without casting
• Misunderstanding path syntax
• Assuming jsonb_set can't update nested keys

5. You have a JSONB column data storing user info. You want to increment the "visits" count inside "stats" by 1, but visits might be missing. Which query correctly updates data to add or increment visits?

hard

A. UPDATE table SET data = data || '{"stats":{"visits":1}}';

B. UPDATE table SET data = jsonb_set(data, '{stats,visits}', '1');

C. UPDATE table SET data = jsonb_set(data, '{stats,visits}', to_jsonb((COALESCE((data->'stats'->>'visits')::int, 0) + 1)));

D. UPDATE table SET data = jsonb_set(data, '{visits}', to_jsonb(1));

Solution

1. Step 1: Understand the need to increment or add visits

   The query must get the current visits count, convert it to integer, handle missing key with COALESCE to 0, add 1, then convert back to JSONB.

2. Step 2: Analyze each option

   UPDATE table SET data = jsonb_set(data, '{stats,visits}', to_jsonb((COALESCE((data->'stats'->>'visits')::int, 0) + 1))); correctly uses COALESCE and to_jsonb to update visits inside stats. UPDATE table SET data = jsonb_set(data, '{stats,visits}', '1'); sets visits to string '1' always. UPDATE table SET data = data || '{"stats":{"visits":1}}'; overwrites stats with visits=1, losing other keys. UPDATE table SET data = jsonb_set(data, '{visits}', to_jsonb(1)); updates visits at top level, not inside stats.

3. Final Answer:

   UPDATE table SET data = jsonb_set(data, '{stats,visits}', to_jsonb((COALESCE((data->'stats'->>'visits')::int, 0) + 1))); -> Option C

4. Quick Check:

   Use COALESCE and to_jsonb for safe increment [OK]

Hint: Use COALESCE and cast to increment JSONB numbers safely [OK]

Common Mistakes:

• Overwriting nested objects accidentally
• Not casting string to int before increment
• Updating wrong JSONB path