I have an issue with a Query I'm conducting to do a search on a Database of events. The purpose is about sports and the structure is:
id_event event_sport event_city
1 10 153
2 12 270
3 09 135
The table sports is like:
sport_id sport_name
1 Basketball
and the table cities is:
city_id city_name
1 NYC
So things get complicated, because my events table is like:
id_event event_sport event_city
1 10,12 153,270
2 7,14 135,271
3 8,12 143,80
and I have a multi-input search form, so that people can search for events in their city for multiple sports or for multiple cities. I'm using Chosen
The search resultant from Chosen is, for example:
City = 153,270 (if user selected more than one city)
Sport = 12 (if user only selected one sport, can be "9,15")
So what I need is to search for multiple values on cities and sports in the same column, separated by commas, knowing that sometimes we can be searching only for one value, if user didn't input more than one.
My current query is:
SELECT * FROM events e
LEFT JOIN cities c ON e.event_city=c.city_id
LEFT JOIN sports s ON e.event_sport=s.sport_id
WHERE FIND_IN_SET('1CITY', e.event_city) AND FIND_IN_SET('1SPORT', e.event_sport)
;
Which is good to search for one city, but if the user searches for two or more, I don't have way to show it.
When the user inputs multiple cities and/or sports, split it on commas, and then the query should look like:
SELECT * FROM events e
LEFT JOIN cities c on e.event_city = c.city_id
LEFT JOIN sports s ON e.event_sport = s.sport_id
WHERE (FIND_IN_SET('$city[0]', e.event_city) OR FIND_IN_SET('$city[1]', e.event_city) OR ...)
AND (FIND_IN_SET('$sport[0]', e.event_sport) OR FIND_IN_SET('$sport[1]', e.event_sport) OR ...)
Using PHP you can build up those OR
expressions with:
$city_list = implode(' OR ', array_map(function($x) { return "FIND_IN_SET('$x', e.event_city)"; }, explode(',', $_POST['cities'])));
Do the same to make $sport_list, and then your SQL string would contain:
WHERE ($city_list) AND ($sport_list)
As you can see, this is really convoluted and inefficient, I recommend you normalize your schema as suggested in the comments.