Create JSON object from recursive tree structure

Question

Having this simple many-to-many self-referential structure.
An item owns other items through the joins table:

CREATE TABLE items (
 item_id serial PRIMARY KEY
, title text
);
CREATE TABLE joins (
 id serial PRIMARY KEY
, item_id int
, child_id int
);
INSERT INTO items (item_id, title) VALUES
 (1, 'PARENT')
, (2, 'LEVEL 2')
, (3, 'LEVEL 3.1')
, (4, 'LEVEL 4.1')
, (5, 'LEVEL 4.2')
, (6, 'LEVEL 3.2')
;
INSERT INTO joins (item_id, child_id) VALUES
 (1, 2)
, (2, 3)
, (3, 4)
, (3, 5)
, (2, 6)
;

db<>fiddle here

I am trying to retrieve a whole tree structure as JSON for a given item.
For example, to query the item with item_id 1 (pseudo-code):

SELECT i.*, fulltree from items i where item_id = 1;

Desired output for fulltree:

{
 id: 1,
 title: "PARENT",
 children: [
 {
 id: 2,
 title: "LEVEL 2",
 children: [
 {
 id: 3,
 title: "LEVEL 3.1",
 children: [
 {
 id: 4,
 title: "LEVEL 4.1"
 },
 {
 id: 5,
 title: "LEVEL 4.2"
 }
 ]
 },
 {
 id: 6,
 title: "LEVEL 3.2"
 }
 ]
 }
 ]
}

After digging into the JSON capabilities Postgres offers, I managed such output by repeating a nested query. Simple but ugly, and limited to the amount of repeats. :/

I've found out about recursive queries. The examples found here and there are not that simple. It's hard to finding an entrypoint to understanding the technique and adapt it to my needs.

I hope the example here will be simple enough to find help from experienced users.

• Evan Carroll: Does one child ever have two parents?
• Ben: @EvanCarroll Found many examples with setup on single table having a parent_id column (ltree style, or ancestry... don't know how to call it)... and that would have been simpler I guess. So my answer to your question is de facto yes. I'd be glad to understand how it could be done that way, or understand why it should not be done that way. thx for the edit! the link might not be that visible for sure
• Erwin Brandstetter: @ISupportTheBoycott: For the purpose of this question, the number of parents seems to make no difference.

Answer 1

A recursive CTE (rCTE) does not allow aggregation in the recursive term. So there is no simple solution.

I suggest a recursive function for an elegant solution:

CREATE OR REPLACE FUNCTION f_item_tree(_item_id int)
 RETURNS jsonb
 LANGUAGE sql STABLE PARALLEL SAFE AS
$func$
SELECT jsonb_agg(sub)
FROM (
 SELECT i.*, f_item_tree(i.item_id) AS children
 FROM joins j
 JOIN items i ON i.item_id = j.child_id
 WHERE j.item_id = _item_id
 ORDER BY i.item_id
 ) sub
$func$;

fiddle

Bare call:

SELECT to_jsonb(sub) AS tree
FROM (
 SELECT *, f_item_tree(item_id) AS children
 FROM items
 WHERE item_id = 1 -- root item_id HERE
 ) sub;

To strip objects with NULL value (no children) and prettify:

SELECT jsonb_pretty(jsonb_strip_nulls(to_jsonb(sub))) AS tree
FROM (
 SELECT *, f_item_tree(item_id) AS children
 FROM items
 WHERE item_id = 1 -- root item_id HERE
 ) sub;

Produces your desired output exactly (one complete tree):

{
 "title": "PARENT",
 "item_id": 1,
 "children": [
 {
 "title": "LEVEL 2",
 "item_id": 2,
 "children": [
 {
 "title": "LEVEL 3.1",
 "item_id": 3,
 "children": [
 {
 "title": "LEVEL 4.1",
 "item_id": 4
 },
 {
 "title": "LEVEL 4.2",
 "item_id": 5
 }
 ]
 },
 {
 "title": "LEVEL 3.2",
 "item_id": 6
 }
 ]
 }
 ]
}

jsonb_strip_nulls() ...

Deletes all object fields that have null values from the given JSON value, recursively.

If there can be other fields with null values that you want to keep, you have to do more. Like:

CREATE OR REPLACE FUNCTION f_item_tree(_item_id int)
 RETURNS jsonb
 LANGUAGE sql STABLE PARALLEL SAFE AS
$func$
SELECT jsonb_agg(
 CASE WHEN children IS NULL
 THEN to_jsonb(sub) - 'children'
 -- THEN jsonb_build_object('title', title, 'item_id', item_id) -- alt: spell out
 ELSE to_jsonb(sub) END
 )
FROM (
 SELECT i.*, f_item_tree(i.item_id) AS children
 FROM joins j
 JOIN items i ON i.item_id = j.child_id
 WHERE j.item_id = _item_id
 ORDER BY i.item_id
 ) sub
$func$;

fiddle

Later, closely related answer with alternatives (most notably a maximum recursion level):

• How to turn a set of flat trees into a single tree with multiple leaves?

Answer 2

Here is an example query,

WITH RECURSIVE t(item_id, json) AS (
 SELECT item_id, to_jsonb(items)
 FROM items
 WHERE NOT EXISTS (
 SELECT 1
 FROM joins
 WHERE items.item_id = joins.item_id
 )
 UNION ALL
 SELECT parent.item_id, to_jsonb(parent) || jsonb_build_object( 'children', t.json )
 FROM t
 JOIN joins AS j
 ON t.item_id = j.child_id
 JOIN items AS parent
 ON j.item_id = parent.item_id
)
SELECT item_id, jsonb_pretty(json)
FROM t
WHERE item_id = 1;

 item_id | jsonb_pretty 
---------+---------------------------------------
 1 | { +
 | "title": "PARENT", +
 | "item_id": 1, +
 | "children": { +
 | "title": "LEVEL 2", +
 | "item_id": 2, +
 | "children": { +
 | "title": "LEVEL 3.2", +
 | "item_id": 6 +
 | } +
 | } +
 | }
 1 | { +
 | "title": "PARENT", +
 | "item_id": 1, +
 | "children": { +
 | "title": "LEVEL 2", +
 | "item_id": 2, +
 | "children": { +
 | "title": "LEVEL 3.1", +
 | "item_id": 3, +
 | "children": { +
 | "title": "LEVEL 4.1",+
 | "item_id": 4 +
 | } +
 | } +
 | } +
 | }
 1 | { +
 | "title": "PARENT", +
 | "item_id": 1, +
 | "children": { +
 | "title": "LEVEL 2", +
 | "item_id": 2, +
 | "children": { +
 | "title": "LEVEL 3.1", +
 | "item_id": 3, +
 | "children": { +
 | "title": "LEVEL 4.2",+
 | "item_id": 5 +
 | } +
 | } +
 | } +
 | }
(3 rows)

Note, we're not actually merging the paths to form a completed tree. You either have to build the tree from the root node down, or the leaf nodes to the top. In this case, you'll have to merge the discrete paths. Look for a deep json merge in Javascript and tie it together with plv8.