Tengo la siguiente tabla:

+----+---------------------+-------------+-----------------+
| id | stat_time           |       reads |          writes |
+----+---------------------+-------------+-----------------+
|  1 | 2013-07-18 20:00:00 |    42614543 |         1342129 |
|  2 | 2013-07-18 21:00:00 |    23085319 |          326139 |
|  3 | 2013-07-25 12:00:00 |           0 |           39639 |
|  4 | 2013-07-25 13:00:00 |      754166 |           39639 |
|  5 | 2013-07-25 14:00:00 |      693382 |          295323 |
|  6 | 2013-07-25 15:00:00 |     1334462 |               0 |
|  7 | 2013-07-25 16:00:00 |    10748367 |          261489 |
|  9 | 2013-07-25 17:00:00 |     4337294 |               0 |
| 10 | 2013-07-25 18:00:00 |     3002796 |               0 |
| 11 | 2013-07-25 20:00:00 |     3002832 |               0 |
| 12 | 2013-07-25 23:00:00 |           0 |          333468 |
| 13 | 2013-07-26 17:00:00 |    10009585 |               0 |
| 15 | 2013-07-26 18:00:00 |     6005752 |               0 |
| 17 | 2013-07-26 21:00:00 |      333663 |               0 |
+----+---------------------+-------------+-----------------+

Me gustaría realizar algo como esto:

SELECT stat_time, SUM(reads), SUM(writes) from this_table GROUP BY stat_time HAVING 'the same day'..

Entonces, un comando que genera tres filas (una para 2013-07-18, la segunda para 2013-07-25 y la tercera para 2013-07-26) y en cada fila de la columna de lecturas / escrituras hay una suma de Lee / escribe en este día.

Gracias David