How do you insert into an SQLite JSON array without duplicates? (like a set)

Question

Given this table:

CREATE TABLE "carts" (
    "id"    INTEGER NOT NULL,
    "products"  TEXT NOT NULL,
    PRIMARY KEY("id" AUTOINCREMENT)
)

Where the products column contains text values representing JSON arrays of numbers like [12,13,14] , how can I insert a single item without duplicates?

Examples

Add 17 to [12,13,14] to give [12,13,14,17] .

Add 13 to [12,13,14] to give [12,13,14] (no change therefore duplicate avoided).

Answer 1

WITH
    new_product(id, product_id) AS (VALUES (1, 13)),
    new_records AS (
        SELECT carts.id, json_insert(carts.products, '$[#]', np.product_id) AS products
        FROM carts, new_product AS np
        WHERE carts.id = np.id
          AND NOT (carts.products like         np.product_id || ',%'
                OR carts.products like '%,' || np.product_id
                OR carts.products like '%,' || np.product_id || ',%')
    )
UPDATE carts SET products = new_records.products
FROM new_records
WHERE carts.id = new_records.id;

Answer 2

SQLite does not have a special JSON data type (although you can use it in the definition of a column).
All JSON-like values are strings (data type TEXT ) and you can easily use string functions and the operator LIKE to check for a pattern in their values:

UPDATE carts
SET products = json_insert(products, '$[#]', ?)
WHERE id = 1
  AND REPLACE(REPLACE(products, '[', ','), ']', ',') NOT LIKE '%,' || ? || ',%'; 

Replace ? with the value that you want to insert.

See the demo .

Answer 3

I have found this to be the easiest way:

1. Insert the item into the array using the standard function json_insert .

> SELECT json_insert('[12,13,14]','$[#]',13) AS tempArray
tempArray
[12,13,14,13]

2. Use the table-valued function json_each to break apart the array into a temporary table.

> SELECT * FROM (SELECT json_insert('[12,13,14]','$[#]',13) AS tempArray), json_each(tempArray)
tempArray   key value   type    atom    id  parent  fullkey path
[12,13,14,13]   0   12  integer 12  1       $[0]    $
[12,13,14,13]   1   13  integer 13  2       $[1]    $
[12,13,14,13]   2   14  integer 14  3       $[2]    $
[12,13,14,13]   3   13  integer 13  4       $[3]    $

3. Take only the value column (as the others are not needed).

> SELECT value FROM (SELECT json_insert('[12,13,14]','$[#]',13) AS tempArray), json_each(tempArray)
value
12
13
14
13

4. Use DISTINCT to remove duplicates.

> SELECT DISTINCT value FROM (SELECT json_insert('[12,13,14]','$[#]',13) AS tempArray), json_each(tempArray)
value
12
13
14

5. Use the aggregation function json_group_array to combine the results in a JSON array text value.

> SELECT json_group_array(DISTINCT value) FROM (SELECT json_insert('[12,13,14]','$[#]',13) AS tempArray), json_each(tempArray)
json_group_array(DISTINCT value)
[12,13,14]

6. Stick this statement into an UPDATE statement, replacing the example array with a reference to the desired field.

UPDATE carts
SET product = (SELECT json_group_array(DISTINCT value) FROM (SELECT json_insert(carts.product,'$[#]',13) AS tempArray), json_each(tempArray))
WHERE id = 1