我有这张桌子：
activity - activity_id - (INT) PRIMARY,
item_id (INT),
user_id (INT)
lat (FLOAT),
lng (FLOAT),
created_at (DATETIME)
我想运行以下查询：
SELECT item_id, count(distinct user_id) as c_c
FROM activity
WHERE (lat BETWEEN X and X and lng BETWEEN X and X)
and created_at >= DATE(NOW(), INTERVAL 15 DAY)
GROUP by item_id
ORDER by c_c desc
LIMIT 10
我以为我在以下字段上做了一个索引：
lat,
lng,
created_at,
bid
那 - 它运行得非常快，但它仍然运行缓慢。这里的最佳指数是多少？我是否将user_id包含在那里？
任何建议都将不胜感激。
更新：
这是EXPLAIN：
+----+-------------+----------+-------+-----------------------------------------+---------+---------+------+-------+----------------------------------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+----------+-------+-----------------------------------------+---------+---------+------+-------+----------------------------------------------+
| 1 | SIMPLE | activity | index | lat,created_lat_lng,created_at,lat_user | item_id | 4 | NULL | 51510 | Using where; Using temporary; Using filesort |
+----+-------------+----------+-------+-----------------------------------------+---------+---------+------+-------+----------------------------------------------+
桌上的索引：
+-----------------+----------------------+
| Name | Columns |
+-----------------+----------------------+
| lat | lat, lng |
| created_lat_lng | lat, lng, created_at |
| created_at | created_at |
| lat_user | lat,lng,user_id,item |
| item_id | item_id |
+-----------------+----------------------+