I'm trying to convert a function in Postgres into a select
query which I intend to use as a view. The reason is that I'd like to access it from a client via a select
query with a where
clause instead of using a parameter as with the function.
The table represents a tree (and adjacency list) and is defined as follow:
CREATE TABLE tree (
id serial primary key,
parent_id int references tree(id)
);
INSERT INTO tree (id, parent_id) VALUES
(1,null)
, (2,1), (3,2), (4,3), (5,3)
, (6,5), (7,6), (8,4), (9,8)
, (10,9), (11,9), (12,7)
, (13,12), (14,12), (15,11)
, (16,15), (17,16), (18,14)
, (19,13), (20,19), (21,20);
SELECT setval ('tree_id_seq', 21); -- reset sequence
-- This produces a tree like:
-- +-- <10>
-- /
-- +-- <4> -- <8> --- <9> -+- <11> --- <15> --- <16> --- <17>
-- /
-- <1> --- <2> --- <3> -+
-- \
-- +-- <5> --- <6> --- <7> --- <12> -+- <14> --- <18>
-- \
-- \
-- \
-- \
-- +-- <13> --- <19> --- <20> --- <21>
--
To get a path from any node in the tree to the root in order, I use this function:
create or replace function _tree(rev int)
returns table(id int, parent_id int, depth int) as $$
declare
sql text;
begin
sql = 'WITH RECURSIVE tree_list(id, parent_id, depth) AS (
SELECT id, parent_id, 1 FROM tree WHERE id = ' || rev ||
'UNION
SELECT p.id, p.parent_id, r.depth + 1
FROM tree p, tree_list r
WHERE p.id = r.parent_id
)
SELECT id, parent_id, depth FROM tree_list order by id;';
return query execute sql;
end;
$$ language plpgsql;
A query would look like select * from _tree(15)
. The question is how would I go about converting this function into a view, so I could call select * from tree where id <= 15
. Also, would a view be executed at the same speed as the function (i.e. would the where clause be considered while executing the query)?
First of all, you can simplify your function quite a bit. This simpler SQL function does the same:
CREATE OR REPLACE FUNCTION f_tree(_rev int)
RETURNS TABLE(id int, parent_id int, depth int) AS
$func$
WITH RECURSIVE tree_list AS (
SELECT t.id, t.parent_id, 1 -- AS depth
FROM tree t
WHERE t.id = $1
UNION ALL -- no point using UNION
SELECT t.id, t.parent_id, r.depth + 1
FROM tree_list r
JOIN tree t ON t.id = r.parent_id
)
SELECT t.id, t.parent_id, t.depth
FROM tree_list t
ORDER BY t.id;
$func$ LANGUAGE sql;
Call:
select * from f_tree(15);
You could use plpgsql, might be slightly beneficial for cashing the query plan in versions before PostgreSQL 9.2. But you voided the only theoretical benefit by using dynamic SQL without need. This makes no sense at all. Simplify to plain SQL.
Use UNION ALL
instead of UNION
, cheaper since there cannot be dupes by design.
Obviously, you can replace this with plain SQL:
WITH RECURSIVE tree_list AS (
SELECT t.id, t.parent_id, 1 AS depth
FROM tree t
WHERE t.id = 15 -- enter parameter here
UNION ALL
SELECT t.id, t.parent_id, r.depth + 1
FROM tree_list r
JOIN tree t ON t.id = r.parent_id
)
SELECT t.id, t.parent_id, t.depth
FROM tree_list t
ORDER BY t.id;
Does the same.
Now, the VIEW
is a trivial matter:
CREATE OR REPLACE VIEW v_tree15 AS
WITH RECURSIVE tree_list AS (
SELECT t.id, t.parent_id, 1 AS depth
FROM tree t
WHERE t.id <= 15 -- only detail to change
UNION ALL
SELECT t.id, t.parent_id, r.depth + 1
FROM tree_list r
JOIN tree t ON t.id = r.parent_id
)
SELECT t.id, t.parent_id, t.depth
FROM tree_list t
ORDER BY t.id;
The result does not make a lot of sense to me, but the question does not define anything more sensible ..
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With