Reputation: 7535
I have a database with an Items
table that looks something like this:
id
name
category (int)
There are several hundred thousand records. Each item
can be in one of 7 different categories
, which correspond to a categories
table:
id
category
I want a query that chooses 1 random item, from each category. Whats the best way of approaching that? I know to use Order By rand()
and LIMIT 1
for similar random queries, but I've never done something like this.
Upvotes: 12
Views: 12270
Reputation: 125
Starting from mysql8 you can solve this by using Windows functions,
avoiding the extra join condition with category
table
https://dev.mysql.com/doc/refman/8.0/en/window-functions-usage.html
SELECT * FROM (
SELECT FIRST_VALUE(id) OVER(PARTITION BY category ORDER BY RAND()) AS first_id FROM items
Order BY RAND()) AS r
GROUP BY first_id
Upvotes: 0
Reputation: 31
Change order of the original table (random order), before final select:
select * from
(select category, id, name from categories order by rand()) as tab
group by 1
Upvotes: 0
Reputation: 272146
This query returns all items joined to categories in random order:
SELECT
c.id AS cid, c.category, i.id AS iid, i.name
FROM categories c
INNER JOIN items i ON c.id = i.category
ORDER BY RAND()
To restrict each category to one, wrap the query in a partial GROUP BY
:
SELECT * FROM (
SELECT
c.id AS cid, c.category, i.id AS iid, i.name
FROM categories c
INNER JOIN items i ON c.id = i.category
ORDER BY RAND()
) AS shuffled_items
GROUP BY cid
Note that when a query has both GROUP BY
and ORDER BY
clause, the grouping is performed before sorting. This is why I have used two queries: the first one sorts the results, the second one groups the results.
I understand that this query isn't going to win any race. I am open to suggestions.
Upvotes: 11
Reputation: 4190
Try this
SELECT id, name, category from Items where
(
select count(*) from Items i where i.category = Items.category
GROUP BY i.category ORDER BY rand()
) <= 1
REF: http://www.xaprb.com/blog/2006/12/07/how-to-select-the-firstleastmax-row-per-group-in-sql/
Upvotes: 0
Reputation: 19882
Here is a simple solution. Let suppose you have this table.
id name category
1 A 1
2 B 1
3 C 1
4 D 2
5 E 2
6 F 2
7 G 3
8 H 3
9 I 3
Use this query
select
c.id,
c.category,
(select name from category where category = c.category group by id order by rand() limit 1) as CatName
from category as c
group by category
Upvotes: 0
Reputation: 754
Please note: in the following example I am assuming your table is named "items" not "Items" because you also said the other table was named "categories" (second table name not capitalized).
The SQL for what you want to do would roughly be:
`SELECT items.id AS item_id,
items.name AS item_name,
items.category AS item_category_id,
categories.id AS category_id,
categories.category AS category_name
FROM items, category
WHERE items.category = categories.id
ORDER BY rand()
LIMIT 1`
Upvotes: -1