lemmy/crates/db_schema/replaceable_schema/triggers.sql
dullbananas c4fc3a8ede
Optimize stuff in attempt to fix high amount of locks, and fix comment_aggregates.child_count (#4696)
* separate triggers

* auto_explain.log_triggers=on

* Revert "auto_explain.log_triggers=on"

This reverts commit 078b2dbb9bbb878078f1089cffda06b6458d127c.

* Revert "separate triggers"

This reverts commit 95600da4afb49e1dc203c2d61da4c57858e107db.

* bring back migration

* re-order statements

* add comment about statement ordering

* no redundant updates

* optimize post_aggregates update in comment trigger

* set comment path in trigger

* update comment_aggregates.child_count using trigger

* move `LEFT JOIN post` to inner query

* clean up newest_comment_time_necro

* add down.sql
2024-05-09 08:18:55 -04:00

570 lines
17 KiB
PL/PgSQL

-- A trigger is associated with a table instead of a schema, so they can't be in the `r` schema. This is
-- okay if the function specified after `EXECUTE FUNCTION` is in `r`, since dropping the function drops the trigger.
--
-- Tables that are updated by triggers should not have foreign keys that aren't set to `INITIALLY DEFERRED`
-- (even if only other columns are updated) because triggers can run after the deletion of referenced rows and
-- before the automatic deletion of the row that references it. This is not a problem for insert or delete.
--
-- After a row update begins, a concurrent update on the same row can't begin until the whole
-- transaction that contains the first update is finished. To reduce this locking, statements in
-- triggers should be ordered based on the likelihood of concurrent writers. For example, updating
-- site_aggregates should be done last because the same row is updated for all local stuff. If
-- it were not last, then the locking period for concurrent writers would extend to include the
-- time consumed by statements that come after.
--
--
-- Create triggers for both post and comments
CREATE FUNCTION r.creator_id_from_post_aggregates (agg post_aggregates)
RETURNS int IMMUTABLE PARALLEL SAFE RETURN agg.creator_id;
CREATE FUNCTION r.creator_id_from_comment_aggregates (agg comment_aggregates)
RETURNS int IMMUTABLE PARALLEL SAFE RETURN (
SELECT
creator_id
FROM
comment
WHERE
comment.id = agg.comment_id LIMIT 1
);
CREATE PROCEDURE r.post_or_comment (table_name text)
LANGUAGE plpgsql
AS $a$
BEGIN
EXECUTE replace($b$
-- When a thing gets a vote, update its aggregates and its creator's aggregates
CALL r.create_triggers ('thing_like', $$
BEGIN
WITH thing_diff AS ( UPDATE
thing_aggregates AS a
SET
score = a.score + diff.upvotes - diff.downvotes, upvotes = a.upvotes + diff.upvotes, downvotes = a.downvotes + diff.downvotes, controversy_rank = r.controversy_rank ((a.upvotes + diff.upvotes)::numeric, (a.downvotes + diff.downvotes)::numeric)
FROM (
SELECT
(thing_like).thing_id, coalesce(sum(count_diff) FILTER (WHERE (thing_like).score = 1), 0) AS upvotes, coalesce(sum(count_diff) FILTER (WHERE (thing_like).score != 1), 0) AS downvotes FROM select_old_and_new_rows AS old_and_new_rows GROUP BY (thing_like).thing_id) AS diff
WHERE
a.thing_id = diff.thing_id
AND (diff.upvotes, diff.downvotes) != (0, 0)
RETURNING
r.creator_id_from_thing_aggregates (a.*) AS creator_id, diff.upvotes - diff.downvotes AS score)
UPDATE
person_aggregates AS a
SET
thing_score = a.thing_score + diff.score FROM (
SELECT
creator_id, sum(score) AS score FROM thing_diff GROUP BY creator_id) AS diff
WHERE
a.person_id = diff.creator_id
AND diff.score != 0;
RETURN NULL;
END;
$$);
$b$,
'thing',
table_name);
END;
$a$;
CALL r.post_or_comment ('post');
CALL r.post_or_comment ('comment');
-- Create triggers that update counts in parent aggregates
CREATE FUNCTION r.parent_comment_ids (path ltree)
RETURNS SETOF int
LANGUAGE sql
IMMUTABLE parallel safe
BEGIN
ATOMIC
SELECT
comment_id::int
FROM
string_to_table (ltree2text (path), '.') AS comment_id
-- Skip first and last
LIMIT (nlevel (path) - 2) OFFSET 1;
END;
CALL r.create_triggers ('comment', $$
BEGIN
UPDATE
person_aggregates AS a
SET
comment_count = a.comment_count + diff.comment_count
FROM (
SELECT
(comment).creator_id, coalesce(sum(count_diff), 0) AS comment_count
FROM select_old_and_new_rows AS old_and_new_rows
WHERE
r.is_counted (comment)
GROUP BY (comment).creator_id) AS diff
WHERE
a.person_id = diff.creator_id
AND diff.comment_count != 0;
UPDATE
comment_aggregates AS a
SET
child_count = a.child_count + diff.child_count
FROM (
SELECT
parent_id,
coalesce(sum(count_diff), 0) AS child_count
FROM (
-- For each inserted or deleted comment, this outputs 1 row for each parent comment.
-- For example, this:
--
-- count_diff | (comment).path
-- ------------+----------------
-- 1 | 0.5.6.7
-- 1 | 0.5.6.7.8
--
-- becomes this:
--
-- count_diff | parent_id
-- ------------+-----------
-- 1 | 5
-- 1 | 6
-- 1 | 5
-- 1 | 6
-- 1 | 7
SELECT
count_diff,
parent_id
FROM
select_old_and_new_rows AS old_and_new_rows,
LATERAL r.parent_comment_ids ((comment).path) AS parent_id) AS expanded_old_and_new_rows
GROUP BY
parent_id) AS diff
WHERE
a.comment_id = diff.parent_id
AND diff.child_count != 0;
WITH post_diff AS (
UPDATE
post_aggregates AS a
SET
comments = a.comments + diff.comments,
newest_comment_time = GREATEST (a.newest_comment_time, diff.newest_comment_time),
newest_comment_time_necro = GREATEST (a.newest_comment_time_necro, diff.newest_comment_time_necro)
FROM (
SELECT
post.id AS post_id,
coalesce(sum(count_diff), 0) AS comments,
-- Old rows are excluded using `count_diff = 1`
max((comment).published) FILTER (WHERE count_diff = 1) AS newest_comment_time,
max((comment).published) FILTER (WHERE count_diff = 1
-- Ignore comments from the post's creator
AND post.creator_id != (comment).creator_id
-- Ignore comments on old posts
AND post.published > ((comment).published - '2 days'::interval)) AS newest_comment_time_necro,
r.is_counted (post.*) AS include_in_community_aggregates
FROM
select_old_and_new_rows AS old_and_new_rows
LEFT JOIN post ON post.id = (comment).post_id
WHERE
r.is_counted (comment)
GROUP BY
post.id) AS diff
WHERE
a.post_id = diff.post_id
AND (diff.comments,
GREATEST (a.newest_comment_time, diff.newest_comment_time),
GREATEST (a.newest_comment_time_necro, diff.newest_comment_time_necro)) != (0,
a.newest_comment_time,
a.newest_comment_time_necro)
RETURNING
a.community_id,
diff.comments,
diff.include_in_community_aggregates)
UPDATE
community_aggregates AS a
SET
comments = a.comments + diff.comments
FROM (
SELECT
community_id,
sum(comments) AS comments
FROM
post_diff
WHERE
post_diff.include_in_community_aggregates
GROUP BY
community_id) AS diff
WHERE
a.community_id = diff.community_id
AND diff.comments != 0;
UPDATE
site_aggregates AS a
SET
comments = a.comments + diff.comments
FROM (
SELECT
coalesce(sum(count_diff), 0) AS comments
FROM
select_old_and_new_rows AS old_and_new_rows
WHERE
r.is_counted (comment)
AND (comment).local) AS diff
WHERE
diff.comments != 0;
RETURN NULL;
END;
$$);
CALL r.create_triggers ('post', $$
BEGIN
UPDATE
person_aggregates AS a
SET
post_count = a.post_count + diff.post_count
FROM (
SELECT
(post).creator_id, coalesce(sum(count_diff), 0) AS post_count
FROM select_old_and_new_rows AS old_and_new_rows
WHERE
r.is_counted (post)
GROUP BY (post).creator_id) AS diff
WHERE
a.person_id = diff.creator_id
AND diff.post_count != 0;
UPDATE
community_aggregates AS a
SET
posts = a.posts + diff.posts
FROM (
SELECT
(post).community_id,
coalesce(sum(count_diff), 0) AS posts
FROM
select_old_and_new_rows AS old_and_new_rows
WHERE
r.is_counted (post)
GROUP BY
(post).community_id) AS diff
WHERE
a.community_id = diff.community_id
AND diff.posts != 0;
UPDATE
site_aggregates AS a
SET
posts = a.posts + diff.posts
FROM (
SELECT
coalesce(sum(count_diff), 0) AS posts
FROM
select_old_and_new_rows AS old_and_new_rows
WHERE
r.is_counted (post)
AND (post).local) AS diff
WHERE
diff.posts != 0;
RETURN NULL;
END;
$$);
CALL r.create_triggers ('community', $$
BEGIN
UPDATE
site_aggregates AS a
SET
communities = a.communities + diff.communities
FROM (
SELECT
coalesce(sum(count_diff), 0) AS communities
FROM select_old_and_new_rows AS old_and_new_rows
WHERE
r.is_counted (community)
AND (community).local) AS diff
WHERE
diff.communities != 0;
RETURN NULL;
END;
$$);
CALL r.create_triggers ('person', $$
BEGIN
UPDATE
site_aggregates AS a
SET
users = a.users + diff.users
FROM (
SELECT
coalesce(sum(count_diff), 0) AS users
FROM select_old_and_new_rows AS old_and_new_rows
WHERE (person).local) AS diff
WHERE
diff.users != 0;
RETURN NULL;
END;
$$);
-- For community_aggregates.comments, don't include comments of deleted or removed posts
CREATE FUNCTION r.update_comment_count_from_post ()
RETURNS TRIGGER
LANGUAGE plpgsql
AS $$
BEGIN
UPDATE
community_aggregates AS a
SET
comments = a.comments + diff.comments
FROM (
SELECT
old_post.community_id,
sum((
CASE WHEN r.is_counted (new_post.*) THEN
1
ELSE
-1
END) * post_aggregates.comments) AS comments
FROM
new_post
INNER JOIN old_post ON new_post.id = old_post.id
AND (r.is_counted (new_post.*) != r.is_counted (old_post.*))
INNER JOIN post_aggregates ON post_aggregates.post_id = new_post.id
GROUP BY
old_post.community_id) AS diff
WHERE
a.community_id = diff.community_id
AND diff.comments != 0;
RETURN NULL;
END;
$$;
CREATE TRIGGER comment_count
AFTER UPDATE ON post REFERENCING OLD TABLE AS old_post NEW TABLE AS new_post
FOR EACH STATEMENT
EXECUTE FUNCTION r.update_comment_count_from_post ();
-- Count subscribers for communities.
-- subscribers should be updated only when a local community is followed by a local or remote person.
-- subscribers_local should be updated only when a local person follows a local or remote community.
CALL r.create_triggers ('community_follower', $$
BEGIN
UPDATE
community_aggregates AS a
SET
subscribers = a.subscribers + diff.subscribers, subscribers_local = a.subscribers_local + diff.subscribers_local
FROM (
SELECT
(community_follower).community_id, coalesce(sum(count_diff) FILTER (WHERE community.local), 0) AS subscribers, coalesce(sum(count_diff) FILTER (WHERE person.local), 0) AS subscribers_local
FROM select_old_and_new_rows AS old_and_new_rows
LEFT JOIN community ON community.id = (community_follower).community_id
LEFT JOIN person ON person.id = (community_follower).person_id GROUP BY (community_follower).community_id) AS diff
WHERE
a.community_id = diff.community_id
AND (diff.subscribers, diff.subscribers_local) != (0, 0);
RETURN NULL;
END;
$$);
-- These triggers create and update rows in each aggregates table to match its associated table's rows.
-- Deleting rows and updating IDs are already handled by `CASCADE` in foreign key constraints.
CREATE FUNCTION r.comment_aggregates_from_comment ()
RETURNS TRIGGER
LANGUAGE plpgsql
AS $$
BEGIN
INSERT INTO comment_aggregates (comment_id, published)
SELECT
id,
published
FROM
new_comment;
RETURN NULL;
END;
$$;
CREATE TRIGGER aggregates
AFTER INSERT ON comment REFERENCING NEW TABLE AS new_comment
FOR EACH STATEMENT
EXECUTE FUNCTION r.comment_aggregates_from_comment ();
CREATE FUNCTION r.community_aggregates_from_community ()
RETURNS TRIGGER
LANGUAGE plpgsql
AS $$
BEGIN
INSERT INTO community_aggregates (community_id, published)
SELECT
id,
published
FROM
new_community;
RETURN NULL;
END;
$$;
CREATE TRIGGER aggregates
AFTER INSERT ON community REFERENCING NEW TABLE AS new_community
FOR EACH STATEMENT
EXECUTE FUNCTION r.community_aggregates_from_community ();
CREATE FUNCTION r.person_aggregates_from_person ()
RETURNS TRIGGER
LANGUAGE plpgsql
AS $$
BEGIN
INSERT INTO person_aggregates (person_id)
SELECT
id
FROM
new_person;
RETURN NULL;
END;
$$;
CREATE TRIGGER aggregates
AFTER INSERT ON person REFERENCING NEW TABLE AS new_person
FOR EACH STATEMENT
EXECUTE FUNCTION r.person_aggregates_from_person ();
CREATE FUNCTION r.post_aggregates_from_post ()
RETURNS TRIGGER
LANGUAGE plpgsql
AS $$
BEGIN
INSERT INTO post_aggregates (post_id, published, newest_comment_time, newest_comment_time_necro, community_id, creator_id, instance_id, featured_community, featured_local)
SELECT
new_post.id,
new_post.published,
new_post.published,
new_post.published,
new_post.community_id,
new_post.creator_id,
community.instance_id,
new_post.featured_community,
new_post.featured_local
FROM
new_post
INNER JOIN community ON community.id = new_post.community_id;
RETURN NULL;
END;
$$;
CREATE TRIGGER aggregates
AFTER INSERT ON post REFERENCING NEW TABLE AS new_post
FOR EACH STATEMENT
EXECUTE FUNCTION r.post_aggregates_from_post ();
CREATE FUNCTION r.post_aggregates_from_post_update ()
RETURNS TRIGGER
LANGUAGE plpgsql
AS $$
BEGIN
UPDATE
post_aggregates
SET
featured_community = new_post.featured_community,
featured_local = new_post.featured_local
FROM
new_post
INNER JOIN old_post ON old_post.id = new_post.id
AND (old_post.featured_community,
old_post.featured_local) != (new_post.featured_community,
old_post.featured_local)
WHERE
post_aggregates.post_id = new_post.id;
RETURN NULL;
END;
$$;
CREATE TRIGGER aggregates_update
AFTER UPDATE ON post REFERENCING OLD TABLE AS old_post NEW TABLE AS new_post
FOR EACH STATEMENT
EXECUTE FUNCTION r.post_aggregates_from_post_update ();
CREATE FUNCTION r.site_aggregates_from_site ()
RETURNS TRIGGER
LANGUAGE plpgsql
AS $$
BEGIN
-- only 1 row can be in site_aggregates because of the index idx_site_aggregates_1_row_only.
-- we only ever want to have a single value in site_aggregate because the site_aggregate triggers update all rows in that table.
-- a cleaner check would be to insert it for the local_site but that would break assumptions at least in the tests
INSERT INTO site_aggregates (site_id)
VALUES (NEW.id)
ON CONFLICT ((TRUE))
DO NOTHING;
RETURN NULL;
END;
$$;
CREATE TRIGGER aggregates
AFTER INSERT ON site
FOR EACH ROW
EXECUTE FUNCTION r.site_aggregates_from_site ();
-- Change the order of some cascading deletions to make deletion triggers run before the deletion of rows that the triggers need to read
CREATE FUNCTION r.delete_comments_before_post ()
RETURNS TRIGGER
LANGUAGE plpgsql
AS $$
BEGIN
DELETE FROM comment AS c
WHERE c.post_id = OLD.id;
RETURN OLD;
END;
$$;
CREATE TRIGGER delete_comments
BEFORE DELETE ON post
FOR EACH ROW
EXECUTE FUNCTION r.delete_comments_before_post ();
CREATE FUNCTION r.delete_follow_before_person ()
RETURNS TRIGGER
LANGUAGE plpgsql
AS $$
BEGIN
DELETE FROM community_follower AS c
WHERE c.person_id = OLD.id;
RETURN OLD;
END;
$$;
CREATE TRIGGER delete_follow
BEFORE DELETE ON person
FOR EACH ROW
EXECUTE FUNCTION r.delete_follow_before_person ();
-- Triggers that change values before insert or update
CREATE FUNCTION r.comment_change_values ()
RETURNS TRIGGER
LANGUAGE plpgsql
AS $$
DECLARE
id text = NEW.id::text;
BEGIN
-- Make `path` end with `id` if it doesn't already
IF NOT (NEW.path ~ ('*.' || id)::lquery) THEN
NEW.path = NEW.path || id;
END IF;
RETURN NEW;
END
$$;
CREATE TRIGGER change_values
BEFORE INSERT OR UPDATE ON comment
FOR EACH ROW
EXECUTE FUNCTION r.comment_change_values ();