How to properly optimize the work of php and mysql with two databases?

Question

PHP

I am writing a billing system for a small Internet provider. Due to certain reasons, processing of subscribers requires data from two databases at once.
I did like this:

• I make a request to the first base, get 16000 rows and write them to an array:
  

  array(16000) {
    '0689182' =>
    array(19) {
      'id' =>
      string(5) "25491"
      'agrm_id' =>
      ...
      'vg_id' =>
      string(8) "1323"
      'rec' =>
      string(1) "0"
    }
  ...
  }

  
  
• I go through the entire array and write unique id to a string, so that later I can make an exact selection from the second database;
  

  foreach($basesData as $user) {
              $strAgrmId .= $user['agrm_id'] . ',';
          }

  
  
• I make a request to the second base, and I get 16000 rows with additional information:
  

  $sql = "SELECT
                  number,
                  balance,
                  credit
              FROM
                agreements
              WHERE agrm_id IN(" . substr($strAgrmId, 0, -1) . ")";

  
  

  array(16000) {
    '0689182' =>
    array(19) {
      'balance' =>
      string(5) "25491"
      'credit' => 
      string(1) "0"
    }
  ...
  }

  
  
• I unite;
  

  foreach($basesData as $key => $user) {
              $users[$key] = array_merge($basesData[$key], $agrmData[$key]);
          }

  
  

Further along the algorithm, I do a similar manipulation with the selection of information to determine the status and again iterate over the entire array of users:

foreach($users as $key => $user) {

            ..code..

            foreach($arrayPool as $one) {
                        ..code..
                        $users[$key]['status'] = $one['status'];
            }
        }

Nested foreach works with arrays of length 20-30. But as a result, it takes a lot of time to process and collect all the necessary information into one array. On my PC, the script does not complete in 15 minutes.
From here the first question : how to properly organize optimization in such a case?
And second : I have a suspicion that this algorithm is complete nonsense. And I think that it is possible to assemble the same information but using a different algorithm.

• I get the first array with 16000 entries;
• I start iterating through the array and at the same time for each of the users I make a separate request to search for additional information, then I make another request to search and process the status;
• I add the information found for each individual user and continue the algorithm.

Thanks in advance!

Answer 1

You can also access two databases in one query if the user has access - dbName.dbTable.dbField
stackoverflow.com/questions/4076562/query-with-joi...

Answer 2

I don’t know your level of knowledge, so just in case I’ll ask again: exactly two different databases or just tables? These two things are sometimes confused.
Profile where you are actually wasting your time. The current zend engine has a rather strange memory management, which will be thoroughly optimized in php7, but merging only 16 thousand records is a garbage question as long as there is enough memory.
You can somehow twist the processing logic upside down - but for this you need to understand what needs to be done in the end (at the same time, it will be lazy to read long essays).