magento-engcom/import-export-improvements#43: make the customer impor…

…t work with large sets of customer data in the database on low memory systems

 - replace the usage of the CollectionByPagesIteratorFactory with a fetchAll when loading current customers from the database,
   this was causing issues because of the extra overhead that the the iterator caused by using objects to store each row's data
   in this way we can use a similar approach to how the product import loads old skus. See: Magento\Catalog\Model\ResourceModel\Product::getProductEntitiesInfo

app/code/Magento/CustomerImportExport/Model/ResourceModel/Import/Customer/Storage.php

@@ -76,16 +76,17 @@ public function __construct(
     public function load()
     {
         if ($this->_isCollectionLoaded == false) {
-            $collection = clone $this->_customerCollection;
-            $collection->removeAttributeToSelect();
-            $tableName = $collection->getResource()->getEntityTable();
-            $collection->getSelect()->from($tableName, ['entity_id', 'website_id', 'email']);
-
-            $this->_byPagesIterator->iterate(
-                $this->_customerCollection,
-                $this->_pageSize,
-                [[$this, 'addCustomer']]
-            );
+            $connection = $this->_customerCollection->getConnection();
+            $select = $connection->select();
+            $select->from($this->_customerCollection->getMainTable(), ['entity_id', 'website_id', 'email']);
+            $results = $connection->fetchAll($select);
+            foreach ($results as $customer) {
+                $email = strtolower(trim($customer['email']));
+                if (!isset($this->_customerIds[$email])) {
+                    $this->_customerIds[$email] = [];
+                }
+                $this->_customerIds[$email][$customer['website_id']] = $customer['entity_id'];
+            }
 
             $this->_isCollectionLoaded = true;
         }