Here is an example:

$my_array1 = array(
    array(
         ‘id’ => ’1′,
         ‘name’ => ‘tom’
    ),
    array(
         ‘id’ => ’2′,
         ‘name’ => ‘harry’
    )
);
$my_array2 = array(
    array(
         ‘id’ => ’5′,
         ‘name’ => ‘dick’
    ),
    array(
         ‘id’ => ’4′,
         ‘name’ => ‘bob’
    )
);
$merged_array = array_merge( $my_array1 , $my_array2 );

Now, this part worked fine and dandy, except when I went to sort the merged array. I was using usort and a function I had found on the internet:

function arfsort( $a, $fl ){
 $GLOBALS['__ARFSORT_LIST__'] = $fl;
 usort( $a, ‘arfsort_func’ );
 return $a;
}// extended to allow sort direction per field sorted against
function arfsort_func( $a, $b ){
 foreach( $GLOBALS['__ARFSORT_LIST__'] as $f ) {
  switch ($f[1]) { // switch on ascending or descending value
   case “d”:
    $strc = strcmp( strtolower($b[$f[0]]), strtolower($a[$f[0]]) );
    if ( $strc != 0 ){
     return $strc;
    }
   break;
   default:
    $strc = strcmp( strtolower($a[$f[0]]), strtolower($b[$f[0]]) );
    if ( $strc != 0 ){
     return $strc;
    }
   break;
  }
 }
 return 0;
}

What was happening was that the usort was not working. So, I started to think about ways to get around the problem of the usort. My first thoughts were some way to work around it in PHP, so I went and asked a colleague if she new of any way to sort associative arrays in the fashion I needed to. Her first comment was to try and do something in the SQL query before you ever get the data. We talked for a bit and I got this bright idea. I could use the handy IN operator from MySQL. First, I would have to build a string of IDs for the database, then use the IN operator to query for just those IDs in the database. For example:

$my_array1 …. (like above)
$my_array2 …. (like above)

$merged_array = array_merge( $my_array1 , $my_array2 );

foreach($merged_array as $m){
    $merge_list[] = $m['id'];
}

$list_for_query = implode(‘,’,$merge_list);

$query = “SELECT * FROM my_table WHERE id IN ($merge_list) ORDER BY name”;

So, to explain the above code it goes something like this:

1. Make the arrays (mine were two different queries that I needed to merge together)
• Note: the arrays just need to have the row ID in them. No sense in wasting time getting all the columns from the table
2. Merge the arrays (they need to be identical)
3. Make the associative arrays a one level array
4. Implode the list so there are commas separating the ID values
5. Query using those ID values and use ORDER BY to get the correct sorting

Lo and behold, it works. It does just what I want, and I don’t have to fuss about using a custom usort. Hopefully this will help someone else down the road so you don’t waste your time like I did.

ÂÅ

MySQL Replication Documentation:
http://dev.mysql.com/doc/refman/4.1/en/replication-howto.html

One thing to note when setting up replication is that you must have the master and the slave on the same version. Other than that, if you follow the steps in the how-to you’ll be golden.