How to join on subqueries using ARel? - sql

I have a few massive SQL request involving join across various models in my rails application.
A single request can involve 6 to 10 tables.
To run the request faster I want to use sub-queries in the joins (that way I can filter these tables before the join and reduce the columns to the ones I need). I'm trying to achieve this using ARel.
I thought I found the solution to my problem there: How to do joins on subqueries in AREL within Rails,
but things must have changed because I get undefined method '[]' for Arel::SelectManager.
Does anybody have any idea how to achieve this (without using strings) ?

Pierre, I thought a better solution could be the following (inspiration from this gist):
a = A.arel_table
b = B.arel_table
subquery = b.project(b[:a_id].as('A_id')).where{c > 4}
subquery = subquery.as('intm_table')
query = A.join(subquery).on(subquery[:A_id].eq(a[:id]))
No particular reason for naming the alias as "intm_table", I just thought it would be less confusing.

OK so my main problem was that you can't join a Arel::SelectManager ... BUT you can join a table aliasing.
So to generate the request in my comment above:
a = A.arel_table
b = B.arel_table
subquery = B.select(:a_id).where{c > 4}
query = A.join(subquery.as('B')).on(b[:a_id].eq(a[:id])
query.to_sql # SELECT A.* INNER JOIN (SELECT B.a_id FROM B WHERE B.c > 4) B ON A.id = B.a_id

Was looking for this, and was helped by the other answers, but there are some error in both, e.g. A.join(... should be a.join(....
And I also missed how to build an ActiveRecord::Relation.
Here is how to build an ActiveRecord::Relation, in Rails 4
a = A.arel_table
b = B.arel_table
subsel = B.select(b[:a_id]).where(b[:c].gt('4')).as('sub_select')
joins = a.join(subsel).on(subsel[:a_id].eq(a[:id])).join_sources
rel = A.joins(joins)
rel.to_sql
#=> "SELECT `a`.* FROM `a` INNER JOIN (SELECT `b`.`a_id` FROM `b` WHERE (`b`.`c` > 4)) sub_select ON sub_select.`a_id` = `a`.`id`"

Related

Ibis Impala JOIN problem with relabel/name 'column AS newName'

When you use the Ibis API to query impala, for some reason Ibis API forces it to become a subquery (when you join 4-5 tables it suddenly becomes super slow). It simply won't join normally, due to column name overlap problem on joins. I want a way to quickly rename the columns perhaps, isn't that's how SQL usually works?
i0 = impCon.table('shop_inventory')
s0 = impCon.table('shop_expenditure')
s0 = s0.relabel({'element_date': 'spend_element_date', 'element_shop_item': 'spend_shop_item'})
jn = i0.inner_join(s0, [i0['element_date'] == s0['spend_element_date'], i0['element_shop_item'] == s0['spend_shop_item']])
jn.materialize()
jn.execute(limit=900)
Then you have IBIS generating SQL that is SUBQUERYING it without me suggesting it:
SELECT *
FROM (
SELECT `element_date`, `element_shop_item`, `element_address`, `element_expiration`,
`element_category`, `element_description`
FROM dbp.`shop_inventory`
) t0
INNER JOIN (
SELECT `element_shop_item` AS `spend_shop_item`, `element_comm` AS `spend_comm`,
`element_date` AS `spend_date`, `element_amount`,
`element_spend_type`, `element_shop_item_desc`
FROM dbp.`shop_spend`
) t1
ON (`element_shop_item` = t1.`spend_shop_item`) AND
(`element_category` = t1.`spend_category`) AND
(`element_subcategory` = t1.`spend_subcategory`) AND
(`element_comm` = t1.`spend_comm`) AND
(`element_date` = t1.`spend_date`)
LIMIT 900
Why is this so difficult?
It should be ideally as simple as:
jn = i0.inner_join(s0, [s0['element_date'].as('spend_date') == i0['element_date']]
to generate a single: SELECT s0.element_date as spend_date, i0.element_date INNER JOIN s0 dbp.shop_spend ON s0.spend_date == i0.element_date
right?
Are we not ever allowed to have same column names on tables that are being joined? I am pretty sure in raw SQL you can just use "X AS Y" without having to need subquery.
I spent the last few hours struggling with this same issue. A better solution I found is to do the following. Join keeping the variable names the same. Then, before you materialize, only select a subset of the variables such that there isn't any overlap.
So in your code it would look something like this:
jn = i0.inner_join(s0, [i0['element_date'] == s0['element_date'], i0['element_shop_item'] == s0['element_shop_item']])
expr = jn[i0, s0['variable_of_interest_1'],s0['variable_of_interest_2']]
expr.materialize()
See here for more resources
https://docs.ibis-project.org/sql.html

Nested Query Alternatives in AWS Athena

I am running a query that gives a non-overlapping set of first_party_id's - ids that are associated with one third party but not another. This query does not run in Athena, however, giving the error: Correlated queries not yet supported.
Was looking at prestodb docs, https://prestodb.io/docs/current/sql/select.html (Athena is prestodb under the hood), for an alternative to nested queries. The with statement example given doesn't seem to translate well for this not in clause. Wondering what the alternative to a nested query would be - Query below.
SELECT
COUNT(DISTINCT i.third_party_id) AS uniques
FROM
db.ids i
WHERE
i.third_party_type = 'cookie_1'
AND i.first_party_id NOT IN (
SELECT
i.first_party_id
WHERE
i.third_party_id = 'cookie_2'
)
There may be a better way to do this - I would be curious to see it too! One way I can think of would be to use an outer join. (I'm not exactly sure about how your data is structured, so forgive the contrived example, but I hope it would translate ok.) How about this?
with
a as (select *
from (values
(1,'cookie_n',10,'cookie_2'),
(2,'cookie_n',11,'cookie_1'),
(3,'cookie_m',12,'cookie_1'),
(4,'cookie_m',12,'cookie_1'),
(5,'cookie_q',13,'cookie_1'),
(6,'cookie_n',13,'cookie_1'),
(7,'cookie_m',14,'cookie_3')
) as db_ids(first_party_id, first_party_type, third_party_id, third_party_type)
),
b as (select first_party_type
from a where third_party_type = 'cookie_2'),
c as (select a.third_party_id, b.first_party_type as exclude_first_party_type
from a left join b on a.first_party_type = b.first_party_type
where a.third_party_type = 'cookie_1')
select count(distinct third_party_id) from c
where exclude_first_party_type is null;
Hope this helps!
You can use an outer join:
SELECT
COUNT(DISTINCT i.third_party_id) AS uniques
FROM
db.ids a
LEFT JOIN
db.ids b
ON a.first_party_id = b.first_party_id
AND b.third_party_id = 'cookie_2'
WHERE
a.third_party_type = 'cookie_1'
AND b.third_party_id is null -- this line means we select only rows where there is no match
You should also use caution when using NOT IN for subqueries that may return NULL values since the condition will always be true. Your query is comparing a.first_party_id to NULL, which will always be false and so NOT IN will lead to the condition always being true. Nasty little gotcha.
One way to avoid this is to avoid using NOT IN or to add a condition to your subquery i.e. AND third_party_id IS NOT NULL.
See here for a longer explanation.

How to get Django QuerySet 'exclude' to work right?

I have a database that contains schemas for skus, kits, kit_contents, and checklists. Here is a query for "Give me all the SKUs defined for kitcontent records defined for kit records defined in checklist 1":
SELECT DISTINCT s.* FROM skus s
JOIN kit_contents kc ON kc.sku_id = s.id
JOIN kits k ON k.id = kc.kit_id
JOIN checklists c ON k.checklist_id = 1;
I'm using Django, and I mostly really like the ORM because I can express that query by:
skus = SKU.objects.filter(kitcontent__kit__checklist_id=1).distinct()
which is such a slick way to navigate all those foreign keys. Django's ORM produces basically the same as the SQL written above. The trouble is that it's not clear to me how to get all the SKUs not defined for checklist 1. In the SQL query above, I'd do this by replacing the "=" with "!=". But Django's models don't have a not equals operator. You're supposed to use the exclude() method, which one might guess would look like this:
skus = SKU.objects.filter().exclude(kitcontent__kit__checklist_id=1).distinct()
but Django produces this query, which isn't the same thing:
SELECT distinct s.* FROM skus s
WHERE NOT ((skus.id IN
(SELECT kc.sku_id FROM kit_contents kc
INNER JOIN kits k ON (kc.kit_id = k.id)
WHERE (k.checklist_id = 1 AND kc.sku_id IS NOT NULL))
AND skus.id IS NOT NULL))
(I've cleaned up the query for easier reading and comparison.)
I'm a beginner to the Django ORM, and I'd like to use it when possible. Is there a way to get what I want here?
EDIT:
karthikr gave an answer that doesn't work for the same reason the original ORM .exclude() solution doesn't work: a SKU can be in kit_contents in kits that exist on both checklist_id=1 and checklist_id=2. Using the by-hand query I opened my post with, using "checklist_id = 1" produces 34 results, using "checklist_id = 2" produces 53 results, and the following query produces 26 results:
SELECT DISTINCT s.* FROM skus s
JOIN kit_contents kc ON kc.sku_id = s.id
JOIN kits k ON k.id = kc.kit_id
JOIN checklists c ON k.checklist_id = 1
JOIN kit_contents kc2 ON kc2.sku_id = s.id
JOIN kits k2 ON k2.id = kc2.kit_id
JOIN checklists c2 ON k2.checklist_id = 2;
I think this is one reason why people don't seem to find the .exclude() solution a reasonable replacement for some kind of not_equals filter -- the latter allows you to say, succinctly, exactly what you mean. Presumably the former could also allow the query to be expressed, but I increasingly despair of such a solution being simple.
You could do this - get all the objects for checklist 1, and exclude it from the complete list.
sku_ids = skus.values_list('pk', flat=True)
non_checklist_1 = SKU.objects.exclude(pk__in=sku_ids).distinct()

SQL joins vs nested query

This two SQL statements return equal results but first one is much slower than the second:
SELECT leading.email, kstatus.name, contacts.status
FROM clients
JOIN clients_leading ON clients.id_client = clients_leading.id_client
JOIN leading ON clients_leading.id_leading = leading.id_leading
JOIN contacts ON contacts.id_k_p = clients_leading.id_group
JOIN kstatus on contacts.status = kstatus.id_kstatus
WHERE (clients.email = 'some_email' OR clients.email1 = 'some_email')
ORDER BY contacts.date DESC;
SELECT leading.email, kstatus.name, contacts.status
FROM (
SELECT *
FROM clients
WHERE (clients.email = 'some_email' OR clients.email1 = 'some_email')
)
AS clients
JOIN clients_leading ON clients.id_client = clients_leading.id_client
JOIN leading ON clients_leading.id_leading = leading.id_leading
JOIN contacts ON contacts.id_k_p = clients_leading.id_group
JOIN kstatus on contacts.status = kstatus.id_kstatus
ORDER BY contacts.date DESC;
But I'm wondering why is it so? It looks like in the firt statement joins are done first and then WHERE clause is applied and in second is just the opposite. But will it behave the same way on all DB engines (I tested it on MySQL)?
I was expecting DB engine can optimize queries like the fors one and firs apply WHERE clause and then make joins.
There are a lot of different reasons this could be (keying etc), but you can look at the explain mysql command to see how the statements are being executed. If you can run that and if it still is a mystery post it.
You always can replace join with nested query... It's always faster but lot messy...

hibernate - HQL joins on many clauses

I've been reading Hibernate documentation, but I haven't found anything that would explain how to do the following.
I have the following SQL code that I'm trying to convert to HQL:
SELECT {msg.*}, {cmd.*}
FROM Schema.Messages AS msg
LEFT OUTER JOIN schema.send_commands AS cmd
ON cmd.message_key = msg.unique_key
AND ( lower(cmd.status) IN (lower('failed') ) )
WHERE msg.sequence_received < 10";
The mainissue I'm having is that I'm unable to have two clauses on a LEFT OUTER JOIN. HQL allows me to have
ON cmd.message_key = msg.unique_key
, but how do I add the
AND clause 2?
You can add extra join conditions using with keyword, something like this (depends on your mapping):
SELECT m, c
FROM Message m LEFT JOIN m.commands c WITH (lower(c.status) = 'failed')
WHERE m.sequenceReceived < 10
See also:
16.3. Associations and joins