I have a table called users that has a jsonb column called 'history'. This is an array of objects, one of the elements is called uid which is the id of the person visiting the page as follows:
[ {"ip":"...","uid":2} , {"ip":"...","uid":4} , ... ]
I'm running a query that appends the jsonb object with the field uname to make understanding who 'uid' is a bit easier which will produce:
[ {"ip":"...","uid":2,"uname":"bob"} , {"ip":"...","uid":4,"uname":"dave"} , ... ]
I'm currently doing this using the following query (say, where uid=2):
SELECT json_agg(history2||jsonb_build_object('uname',uname::text)) FROM
(SELECT jsonb_array_elements(history) AS history2 FROM users WHERE uid=2) AS table1
LEFT JOIN users AS table2 ON history2->>'uid'=table2.uid
I'm using the subquery to return a table of json objects that's then joined to the user table again to get the username.
My question is: Is there a way of doing this without having the subquery? I've read that lateral joins could be used but all my attempts at this don't seem to work.
Thanks in advance.
You can move jsonb_array_elements into the FROM clause with an outer join:
SELECT jsonb_agg(h.item||jsonb_build_object('uname', u.uname))
FROM users u
LEFT JOIN jsonb_array_elements(u.history) as h(item) on h.item ->> 'uid' = u.uid::text
WHERE u.uid = 2
Related
I am trying to select from one table using the select result from another table. I can run this in two queries but would like to optimize it into just one.
First query.. Select ids where matching other id
select id from lookuptable where paid = '547'
This results in something like this
6316352
6316353
6318409
6318410
6320468
6320469
6320470
6322526
6322527
6324586
6324587
6326648
I would like to then use this result to make another selection. I can do it manually like below. Note, there could be many rows with these values so I've been using a IN statement
select * from "othertable" where id in (6316352,6316353,6318409,6318410,6320468,6320469,6320470,6322526,6322527,6324586,6324587,6326648);
select
ot.*
from
"othertable" as ot
join
lookuptable as lt
on
ot.id = lt.id
where
lt.paid = '547'
The IN operator supports not just value lists but also subqueries, so you can literally write
select * from "othertable" where id in (select id from lookuptable where paid = '547');
I am a frontend developer but I started to write backend stuff. I have spent quite some amount of time trying to figure out how to solve this. I really need some help.
Here are the simplified definitions and relations of two tables:
Relationship between tables
CREATE TABLE IF NOT EXISTS items (
item_id uuid NOT NULL DEFAULT gen_random_uuid() ,
parent_id uuid DEFAULT NULL ,
parent_table parent_tables NOT NULL
);
CREATE TABLE IF NOT EXISTS collections (
collection_id uuid NOT NULL DEFAULT gen_random_uuid() ,
parent_id uuid DEFAULT NULL
);
Our product is an online document collaboration tool, page can have nested pages.
I have a piece of PostgreSQL code for getting all of its ancestor records for given item_ids.
WITH RECURSIVE ancestors AS (
SELECT *
FROM items
WHERE item_id in ( ${itemIds} )
UNION
SELECT i.*
FROM items i
INNER JOIN ancestors a ON a.parent_id = i.item_id
)
SELECT * FROM ancestors
It works fine for nesting regular pages, But if I am going to support nesting collection pages, which means some items' parent_id might refer to "collection" table's collection_id, this code will not work anymore. According to my limited experience, I don't think pure SQL code can solve it. I think writing a PL/pgSQL function might be a solution, but I need to get all ancestor records to given itemIds, which means returning a mix of items and collections records.
So how to return different format of records from a single PL/pgSQL function? I did some research but haven't found any example.
You can make it work by returning a superset as row: comprised of item and collection. One of both will be NULL for each result row.
WITH RECURSIVE ancestors AS (
SELECT 0 AS lvl, i.parent_id, i.parent_table, i AS _item, NULL::collections AS _coll
FROM items i
WHERE item_id IN ( ${itemIds} )
UNION ALL -- !
SELECT lvl + 1, COALESCE(i.parent_id, c.parent_id), COALESCE(i.parent_table, 'i'), i, c
FROM ancestors a
LEFT JOIN items i ON a.parent_table = 'i' AND i.item_id = a.parent_id
LEFT JOIN collections c ON a.parent_table = 'c' AND c.collection_id = a.parent_id
WHERE a.parent_id IS NOT NULL
)
SELECT lvl, _item, _coll
FROM ancestors
-- ORDER BY ?
db<>fiddle here
UNION ALL, not UNION.
Assuming a collection's parent is always an item, while an item can go either way.
We need LEFT JOIN on both potential parent tables to stay in the race.
I added an optional lvl to keep track of the level of hierarchy.
About decomposing row types:
Combine postgres function with query
Record returned from function has columns concatenated
Is there a way to get the matching column name when searching across multiple columns in PostgreSQL?
Say I have the following table structure and query:
CREATE TABLE document (
id serial PRIMARY KEY,
document_content VARCHAR
);
CREATE TABLE story (
id serial PRIMARY KEY,
headline VARCHAR
);
-----
SELECT
"document".*,
story.id,
story.headline
FROM
"document"
INNER JOIN story_document AS Documents_join ON "document".id = Documents_join.document_id
INNER JOIN story ON "story".id = Documents_join.story_id
WHERE to_tsvector(document_content) ## to_tsquery('foo')
OR to_tsvector(headline) ## to_tsquery('foo');
I was thinking of concatenating the value of the two columns, run the full text search, then create a sub query for both columns and re-run the search individually and record the result as a reference, but this would mean executing the search 3x:
SELECT
"document".*,
story.id AS story_id,
story.headline
(SELECT "document".id WHERE to_tsvector(document_content) ## to_tsquery('foo')) AS "matching_document_id",
(SELECT story_id WHERE to_tsvector(headline) ## to_tsquery('foo')) AS "matching_story_id"
FROM
"document"
INNER JOIN story_document AS Documents_join ON "document".id = Documents_join.document_id
RIGHT JOIN story ON "story".id = Documents_join.story_id
WHERE to_tsvector(document_content || ' ' || headline) ## to_tsquery('foo');
How could I get a reference to the column: document_content or headline, where the keyword "foo" was found in one query?
Thanks!
Since the columns are in different tables the best you can do is translate the OR into a UNION:
SELECT
"document".*,
story.id,
story.headline
FROM
"document"
INNER JOIN story_document AS Documents_join ON "document".id = Documents_join.document_id
INNER JOIN story ON "story".id = Documents_join.story_id
WHERE to_tsvector(document_content) ## to_tsquery('foo')
UNION
SELECT
"document".*,
story.id,
story.headline
FROM
"document"
INNER JOIN story_document AS Documents_join ON "document".id = Documents_join.document_id
INNER JOIN story ON "story".id = Documents_join.story_id
WHERE to_tsvector(headline) ## to_tsquery('foo');
Then PostgreSQL doesn't have to build the complete join just to filter out most of the rows. My variant will be fast if the conditions are selective and indexed and you have indexes on the join conditions as well, so that you can get fast nested loop joins.
Here is some more about dealing with OR.
I have two jsonb fields in the table below and I would like to do a query where I filter any key of the dictionary.
My problem is that those dictionaries are inside of a list and when I try to access them with:
SELECT *
FROM public.monitoring_environmentalcontrol
WHERE celery_status->'queue'='0'
I get nothing:
You can use jsonb_array_elements DOC function to achieve your goal plus a LATERAL JOIN 7.2.1.5. LATERAL Subqueries on it:
This is the setup I created:
create table test (
id int,
celery_status jsonb
);
insert into test values
(1,'[{"queue":"a"}, {"queue":"b"}, {"queue":"c"}]'),
(2,'[{"queue":"d"}, {"queue":"e"}, {"queue":"f"}]'),
(3,'[{"queue":"g"}, {"queue":"h"}, {"queue":"i"}]');
This is the query:
select t.id, t.celery_status, obj->>'queue'
from test t
join lateral
jsonb_array_elements(t.celery_status) obj(value) on obj->>'queue' = 'a'
You can see it working here: http://sqlfiddle.com/#!17/bf7bf/6
I have two columns in the same table that I want to join in Postgresql but for some reason I’m getting this error. Don’t know how to figure it out. Please help.
[42P01] ERROR: relation "a" does not exist
Position: 10
X table contains two pools(ABC,XYZ), ids, numbers and description. If an ID exists in one pool but not in the other, it should update description column to “ADD”. Pools need to be joined on number.
UPDATE A
SET A.Description = 'ADD'
FROM X AS A
LEFT JOIN X AS B ON B.number = A.number
AND B.id = 'ABC'
WHERE A.id = 'XYZ'
AND B.number IS NULL
AND A.Description IS NULL;
With standard SQL you can't do a join as part of an update, but what you can do is include a subquery to select the id's to update. The subquery can contain a join. I'm not entirely clear on what you're actually trying to accomplish, but you could do something like this:
UPDATE x SET description='ADD' WHERE number IN (
SELECT a.number FROM x AS a
LEFT OUTER JOIN x AS b ON a.number=b.number AND a.id='XYZ' AND b.id='ABC'
WHERE b.number IS NULL
);
This will join the table x with itself and will select (and update) any numbers's that don't have a matching number in the 'ABC' and 'XYZ' zone.
PostgreSQL does have a UPDATE FROM syntax that does let you update with complex subqueries. It's more flexible but it's non-standard. You can find an example of this type of query here.