Working out the Busiest Day using PHP

Question

I have a table with a timestamp in unixtimestamp for everytime someone logs into the system. I would like to work out the busiest day.

Im a little lost as to where to start, i believe its that you need to work out a count for each day stored into the database (or some kind of string between start and end of the logs?)

Database Structure for logs

Example :

user  | action         | timestamp
admin | User Logged In | 1395257767

User being "Username"

Action being "User Logged In"

Timestamp being Unixtimestamp given at the time they login to the system.

Any help would be a great start

SQLFiddle - http://sqlfiddle.com/#!2/66d4293

Answer 1

Try something like thi.

SELECT COUNT(*) c, DATE(FROM_UNIXTIME( timestamp) ) day FROM log 
GROUP BY  day ORDER BY c DESC LIMIT 1;

DATE(FROM_UNIXTIME( timestamp )) converts to a datetime and then a date.

By the way, you might consider storing your dates as datetime or timestamp structures. Storing as unix timestamps and then having to convert to useable dates is less than ideal.

http://sqlfiddle.com/#!2/a53d1/2

Answer 2

just a quick guess of what you are asking for:

 select date(timestamp),count(*) from logs group by date(FROM_UNIXTIME(timestamp))

example output from a database I tested the query on:

+---------------+----------+
| date(timestamp) | count(*) |
+---------------+----------+
| 2013-12-13    |     4978 |
| 2013-12-14    |     5016 |
| 2013-12-15    |     4999 |
| 2013-12-16    |     5002 |
| 2013-12-17    |     4995 |
| 2013-12-18    |     5000 |
| 2013-12-19    |     5001 |
| 2013-12-20    |     5001 |
| 2013-12-21    |     5001 |
| 2013-12-22    |     4713 |
| 2013-12-30    |     4892 |
| 2013-12-31    |     5064 |
| 2014-01-01    |     5024 |
| 2014-01-02    |     5001 |
| 2014-01-03    |     5002 |
| 2014-01-04    |     5001 |
| 2014-01-05    |     4970 |
| 2014-01-06    |     4989 |
| 2014-01-07    |     4995 |
| 2014-01-08    |     5004 |
| 2014-01-09    |     2531 |

Answer 3

To get the busiest day use DAYOFWEEK

SELECT DAYOFWEEK(timestamp),COUNT(*) FROM logs GROUP BY DAYOFWEEK(timestamp)