How to select unique visitors per date interval?

Question

I've found a few query strings related to what I'm looking for, but nothing that does this. Basically, I have a graph that needs to display unique visitors per time interval (hour, day, month, or year). So, a graph similar to Google Analytics or this Analytics Graph . The table has a visitor ID column and a timestamp column (in the style 0000-00-00 00:00:00). I need to be able to select the unique or total visitors per interval (eg. 32 visitors on the 5th, 30 on the 6th, etc).

I guess my question is, is there anyway to do this efficiently with just a query? Or will I need to use PHP to get all the data (problem with that is that I would have to do a query for each point on the graph, not efficient)?

Answer 1

You will need to aggregate your data. Try this query:

SELECT DATE(data_timestamp), COUNT(visitor_id)
FROM analytics_table
WHERE DATE(data_timestamp) BETWEEN '2011-05-01' AND '2011-05-31'
GROUP BY 1

Here's how you populate your monthly data ( WARNING: UNTESTED! ):

<?php
$sql = "SELECT DATE(data_timestamp), COUNT(visitor_id)
FROM analytics_table
WHERE DATE(data_timestamp) BETWEEN '2011-05-01' AND '2011-05-31'
GROUP BY 1";

$rs = mysql_query($rs);
$date1 = $datex = '2011-05-01';
$date2 = '2011-05-31';

$arrayData = $tmpArray = array();

while( $r = mysql_fetch_array($rs) )
{
   $tmpArray[$r['date']] = $r['count'];
}

while( $datex <= $date2)
{
   if( isset($tmpArray[$datex]) )
   {
      $arrayData[$datex] = $tmpArray[$datex];
   }
   else
   {
      $arrayData[$datex] = 0;
   }
   list( $y, $m, $d) = explode('-', $datex);
   $datex = date('Y-m-d', mktime(0, 0, 0, $m, $d, $y));
}


?>

this query use:

• DATE() to get the date from your data and
• COUNT() to count total data from that particular date.
• and GROUP BY to group your data based on what field you select.

Answer 2

Basically, you can cast the DATETIME/TIMESTAMP into dates, and times separately, and then use a GROUP BY to get the results you need.

MySQL/SQL: Group by date only on a Datetime column

Answer 3

GROUP BY doesn't return 0 value if there is no visitor some day. That will yield to the holes in time continuity. There are two ways:

1) Group like others said and then find holes

2) Create aggregation table with fields "date", "count" etc. and insert there data every day by cron.