How to select distinct values from query and concatenate into one column

Question

I've got a table with a category column. Along with the selection of the clients, I'd like to select on each row the possible values of the category - that is, all the unique values of category in that subset.

My table looks like this:

| id | name          | category    |
------------------------------------
| 1  | Test Client   | Retail      |
| 2  | Test Client 2 | Corporate   |
| 3  | Test Client 3 | Retail      |
| 4  | Test Client 4 | Retail      |
| 5  | Test Client 5 | Leisure     |

I thought GROUP_CONCAT would do the trick:

SELECT `client`.*, GROUP_CONCAT(DISTINCT client.category) AS possible_categories
FROM (`client`)
WHERE  `name`  LIKE '%query%'
GROUP BY `client`.`id`

...but it just gives me that row's category, not the others.

I can do it in code, but it's an O(n) operation and I'd rather save on the processing time. Here's how I could do it in code, for illustrative purposes:

return array_unique(array_map(function($client)
{
    return $client->category;
}, $clients));

The ideal scenario would be to see a table like this:

| id | name          | category    | possible_categories     |
---------------------------------------------------------------
| 1  | Test Client   | Retail      | Retail,Corporate,Leisure |
| 2  | Test Client 2 | Corporate   | Retail,Corporate,Leisure |
| 3  | Test Client 3 | Retail      | Retail,Corporate,Leisure |
| 4  | Test Client 4 | Retail      | Retail,Corporate,Leisure |
| 5  | Test Client 5 | Leisure     | Retail,Corporate,Leisure |

Answer 1

Assuming you mean the possible categories for the matching names:-

SELECT `client`.*, Sub1.possible_categories
FROM (`client`)
CROSS JOIN (SELECT GROUP_CONCAT(DISTINCT client.category) AS possible_categories FROM (`client`) WHERE  `name`  LIKE '%query%') Sub1
WHERE  `name`  LIKE '%query%'

Note the leading wildcard in the LIKEs will probably make it run slowly.

Answer 2

If by "possible categories" you mean all categories, then you need to calculate them separately and add them in:

SELECT `client`.*, cc.possible_categories
FROM `client` cross join 
     (select GROUP_CONCAT(DISTINCT client.category) AS possible_categories
      from `client`
      where `name`  LIKE '%query%'
     ) cc
WHERE  `name`  LIKE '%query%'

Answer 3

I think this would work for you:

SELECT id, name, category, (SELECT GROUP_CONCAT(distinct category) FROM Client) AS possible_categories
FROM client

See the demo on SQLFiddle .