MySQL查询获取联接表结果

Question

There is a chance this has been asked but I'm not entirely sure how to word it which makes searching for an answer a bit difficult.

I have a table of accommodation listings (listingid (primary key), listing_name, description, etc..). I then have another table for attributes relating to each listing (attributeid (primary key), listingid, attribute). Each attribute takes up its own row in the table - I realise this could be much smarter but have inherited the data structure.

My example of what I want to achieve is:

SELECT all listings where attributes have "WiFi" AND "Air Conditioning" AND "Swimming Pool"

Along with these must have attributes, I only want to return listings that match ANY of the following: "Motel", "Hotel", "Apartment".

I was initially doing a INNER JOIN on the attributes table and then going WHERE (attribute = "X" AND attribute = "Y") AND (attribute = "XX" OR attribute = "YY") but realised that the start of this statement will never match multiple rows as it is only ever checking against a single row. So need to consider the entire result of the JOIN (or sub query?) which is where I need to expand my SQL understanding.

I don't really want to go down my typical route of creating x amount of sub-queries as I imagine this will make the entire query extremely slow - especially if a user selected a dozen or more attributes to search on.

Hopefully this explains it enough but let me know if you require more information.

Answer 1

Something like this should work:

select
    acc.*
from
    accomodation acc
    join attribute air on acc.id = air.accomodation_id and air.name='Air'
    join attribute pool on acc.id = pool.accomodation_id and pool.name='Pool'
    and acc.type in ('Motel', 'Hotel', 'Appartment')
;

Answer 2

select l.listingid, l.listing_name 
from listings l join attributes a 
on l.listingid = a.listingid
where a.attribute = 'WiFi' 
and a.attribute = 'Air Conditioning' and a.attribute = 'Swimming Pool'
and l.description in ('Motel', 'Hotel', 'Apartment')

The column names as per your conditions have been assumed. But this should get you what you need.