MultiDimensional Arrays - php

Question

In Codeigniter, I have the following model
function get_item_history($id)
{
//from metadata_history get item_id and corresponding metadata
$this->db->from('metadata_history')->where(array('id'=>$id, 'current_revision'=> "TRUE"));
$query = $this->db->get();
$result = $query->result_array(); //store this in an array
// loop through the array
foreach( $result as $key => $row )
{
$array = array('item_id'=>$row['item_id'], 'current_revision'=> "TRUE");
$this->db->from('history')->where($array);
$query = $this->db->get();
$row['items'] = $query->result_array(); //
$result[$key] = $row;
}
return $result;
}
The problem is that this results in multiple queries to the SQL table increasing the execution time significantly (pagination is not an option)
I want to be able to pass the first query results to the second query as an array, so that I would have only a single go at the database, then rebuild an array from the results.
How should I rewrite this code (the second part)? Will it be faster (I suppose so)?
EDIT
Rebuilding the array from the results is what is flummoxing me.
http://www.phpbuilder.com/board/showthread.php?t=10373847
this is what I probably want, but am failing the jump

Answer 1

You can use inner query here. It is ideal situation for that -
function get_item_history($id)
{
// Here the above requirement can be achieved in a single query.
$sql = "select * from history h
where h.item_id IN (select item_id from metadata_history mh where mh.id = $id
AND mh.current_revision = TRUE) AND h.current_revision = TRUE";
$result = $this->db->query($sql);
//Return whichever column of result you want to return or process result if you want.
$result;
}

Answer 2

You should use JOINs to do this. It'll offload the execution of the query to the server. I can't give you too much more detail without knowing how your database is structured, but check out the docs on JOINs:
http://dev.mysql.com/doc/refman/5.0/en/join.html
http://www.webdesign.org/web-programming/php/mysql-join-tutorial.14876.html
http://www.keithjbrown.co.uk/vworks/mysql/mysql_p5.php

Answer 3

Another option would be to do your wheres in the loop and move the query executation outside of the foreach:
// loop through the array
foreach( $result as $key => $row )
{
$array = array('item_id'=>$row['item_id'], 'current_revision'=> "TRUE");
$this->db->or_where($array);
}
$query = $this->db->get();
$row['items'] = $query->result_array(); //
$result[$key] = $row;

Answer 4

OK this took some work, and I also had to do some adjustments in my view
So the problem can be broken down into two main components
1) Pass the results of the first query as an array to the second one using where_in
2) Reorder/regroup the results of the first array by item_id
My earlier code was doing the second component implicitly
So here is what I did (limits, offsets, ordering have been cut out to improve readablity)
function get_item_history($id)
{
//from metadata_history get item_id and corresponding metadata
$this->db->from('metadata_history')->where(array('id'=>$id, 'current_revision'=> "TRUE"));
$query = $this->db->get();
$result_query1 = $query->result_array(); //store this in an array
foreach ($result_query1 as $key-> $row){
$result[$row['item_id']]['meta_info'] = $row; //the first query contains meta info, that must be passed to the view
$selected_id_array[] = $row['item_id']; //Create a array to pass on to the next query
$result[$row['item_id']]['items'] = array(); //declare an array which will hold the results of second query later
}
$this->db->select('h.*');
$this->db->from('history h');
$this->db->where_in('h.item_id', $selected_id_array);
$this->db->where(array('h.current_revision' => 'TRUE'));
$query = $this->db->get();
$row = $query->result_array();
foreach ($row as $key => $datarow) {
$result[$datarow['item_id']]['items'][] = $datarow; //populate the array we declared earlier with results from second query
}
return $result; // Now this variable holds an array which is indexed by item id and contains the results of second query 'grouped' by item_id
}
So the number of queries have been cut from ~10 to 2.
On my local machine this saves ~50 msec/page, though I am not sure how this will do for larger databases.