CRM/Import/DataSource/CSV.php

   1 <?php
   2 /*
   3  +--------------------------------------------------------------------+
   4  | CiviCRM version 4.6                                                |
   5  +--------------------------------------------------------------------+
   6  | Copyright CiviCRM LLC (c) 2004-2014                                |
   7  +--------------------------------------------------------------------+
   8  | This file is a part of CiviCRM.                                    |
   9  |                                                                    |
  10  | CiviCRM is free software; you can copy, modify, and distribute it  |
  11  | under the terms of the GNU Affero General Public License           |
  12  | Version 3, 19 November 2009.                                       |
  13  |                                                                    |
  14  | CiviCRM is distributed in the hope that it will be useful, but     |
  15  | WITHOUT ANY WARRANTY; without even the implied warranty of         |
  16  | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.               |
  17  | See the GNU Affero General Public License for more details.        |
  18  |                                                                    |
  19  | You should have received a copy of the GNU Affero General Public   |
  20  | License and the CiviCRM Licensing Exception along                  |
  21  | with this program; if not, contact CiviCRM LLC                     |
  22  | at info[AT]civicrm[DOT]org. If you have questions about the        |
  23  | GNU Affero General Public License or the licensing of CiviCRM,     |
  24  | see the CiviCRM license FAQ at http://civicrm.org/licensing        |
  25  +--------------------------------------------------------------------+
  26 */
  27
  28 /**
  29  *
  30  * @package CRM
  31  * @copyright CiviCRM LLC (c) 2004-2014
  32  * $Id$
  33  *
  34  */
  35 class CRM_Import_DataSource_CSV extends CRM_Import_DataSource {
  36   const
  37     NUM_ROWS_TO_INSERT = 100;
  38
  39   /**
  40    * Provides information about the data source
  41    *
  42    * @return array collection of info about this data source
  43    *
  44    *
  45    */
  46   public function getInfo() {
  47     return array('title' => ts('Comma-Separated Values (CSV)'));
  48   }
  49
  50   /**
  51    * Set variables up before form is built
  52    *
  53    */
  54   public function preProcess(&$form) {
  55   }
  56
  57   /**
  58    * This is function is called by the form object to get the DataSource's
  59    * form snippet. It should add all fields necesarry to get the data
  60    * uploaded to the temporary table in the DB.
  61    *
  62    * @param CRM_Core_Form $form
  63    *
  64    * @return void (operates directly on form argument)
  65    */
  66   public function buildQuickForm(&$form) {
  67     $form->add('hidden', 'hidden_dataSource', 'CRM_Import_DataSource_CSV');
  68
  69     $config = CRM_Core_Config::singleton();
  70
  71     $uploadFileSize = CRM_Core_Config_Defaults::formatUnitSize($config->maxFileSize . 'm', TRUE);
  72     $uploadSize = round(($uploadFileSize / (1024 * 1024)), 2);
  73     $form->assign('uploadSize', $uploadSize);
  74     $form->add('File', 'uploadFile', ts('Import Data File'), 'size=30 maxlength=255', TRUE);
  75     $form->setMaxFileSize($uploadFileSize);
  76     $form->addRule('uploadFile', ts('File size should be less than %1 MBytes (%2 bytes)', array(1 => $uploadSize, 2 => $uploadFileSize)), 'maxfilesize', $uploadFileSize);
  77     $form->addRule('uploadFile', ts('Input file must be in CSV format'), 'utf8File');
  78     $form->addRule('uploadFile', ts('A valid file must be uploaded.'), 'uploadedfile');
  79
  80     $form->addElement('checkbox', 'skipColumnHeader', ts('First row contains column headers'));
  81   }
  82
  83   /**
  84    * Process the form submission
  85    *
  86    */
  87   public function postProcess(&$params, &$db, &$form) {
  88     $file = $params['uploadFile']['name'];
  89     $result = self::_CsvToTable($db,
  90       $file,
  91       CRM_Utils_Array::value('skipColumnHeader', $params, FALSE),
  92       CRM_Utils_Array::value('import_table_name', $params),
  93       CRM_Utils_Array::value('fieldSeparator', $params, ',')
  94     );
  95
  96     $form->set('originalColHeader', CRM_Utils_Array::value('original_col_header', $result));
  97
  98     $table = $result['import_table_name'];
  99     $importJob = new CRM_Contact_Import_ImportJob($table);
 100     $form->set('importTableName', $importJob->getTableName());
 101   }
 102
 103   /**
 104    * Create a table that matches the CSV file and populate it with the file's contents
 105    *
 106    * @param object $db
 107    *   Handle to the database connection.
 108    * @param string $file
 109    *   File name to load.
 110    * @param bool $headers
 111    *   Whether the first row contains headers.
 112    * @param string $table
 113    *   Name of table from which data imported.
 114    * @param string $fieldSeparator
 115    *   Character that seperates the various columns in the file.
 116    *
 117    * @return string  name of the created table
 118    */
 119   private static function _CsvToTable(
 120     &$db,
 121     $file,
 122     $headers = FALSE,
 123     $table = NULL,
 124     $fieldSeparator = ','
 125   ) {
 126     $result = array();
 127     $fd = fopen($file, 'r');
 128     if (!$fd) {
 129       CRM_Core_Error::fatal("Could not read $file");
 130     }
 131
 132     $config = CRM_Core_Config::singleton();
 133     // support tab separated
 134     if (strtolower($fieldSeparator) == 'tab' ||
 135       strtolower($fieldSeparator) == '\t'
 136     ) {
 137       $fieldSeparator = "\t";
 138     }
 139
 140     $firstrow = fgetcsv($fd, 0, $fieldSeparator);
 141
 142     // create the column names from the CSV header or as col_0, col_1, etc.
 143     if ($headers) {
 144       //need to get original headers.
 145       $result['original_col_header'] = $firstrow;
 146
 147       $strtolower = function_exists('mb_strtolower') ? 'mb_strtolower' : 'strtolower';
 148       $columns    = array_map($strtolower, $firstrow);
 149       $columns    = str_replace(' ', '_', $columns);
 150       $columns    = preg_replace('/[^a-z_]/', '', $columns);
 151
 152       // need to take care of null as well as duplicate col names.
 153       $duplicateColName = FALSE;
 154       if (count($columns) != count(array_unique($columns))) {
 155         $duplicateColName = TRUE;
 156       }
 157
 158       // need to truncate values per mysql field name length limits
 159       // mysql allows 64, but we need to account for appending colKey
 160       // CRM-9079
 161       foreach ($columns as $colKey => & $colName) {
 162         if (strlen($colName) > 58) {
 163           $colName = substr($colName, 0, 58);
 164         }
 165       }
 166
 167       if (in_array('', $columns) || $duplicateColName) {
 168         foreach ($columns as $colKey => & $colName) {
 169           if (!$colName) {
 170             $colName = "col_$colKey";
 171           }
 172           elseif ($duplicateColName) {
 173             $colName .= "_$colKey";
 174           }
 175         }
 176       }
 177
 178       // CRM-4881: we need to quote column names, as they may be MySQL reserved words
 179       foreach ($columns as & $column) {
 180         $column = "`$column`";
 181       }
 182     }
 183     else {
 184       $columns = array();
 185       foreach ($firstrow as $i => $_) {
 186         $columns[] = "col_$i";
 187       }
 188     }
 189
 190     // FIXME: we should regen this table's name if it exists rather than drop it
 191     if (!$table) {
 192       $table = 'civicrm_import_job_' . md5(uniqid(rand(), TRUE));
 193     }
 194
 195     $db->query("DROP TABLE IF EXISTS $table");
 196
 197     $numColumns = count($columns);
 198     $create = "CREATE TABLE $table (" . implode(' text, ', $columns) . " text) ENGINE=InnoDB DEFAULT CHARACTER SET utf8 COLLATE utf8_unicode_ci";
 199     $db->query($create);
 200
 201     // the proper approach, but some MySQL installs do not have this enabled
 202     // $load = "LOAD DATA LOCAL INFILE '$file' INTO TABLE $table FIELDS TERMINATED BY '$fieldSeparator' OPTIONALLY ENCLOSED BY '\"'";
 203     // if ($headers) {   $load .= ' IGNORE 1 LINES'; }
 204     // $db->query($load);
 205
 206     // parse the CSV line by line and build one big INSERT (while MySQL-escaping the CSV contents)
 207     if (!$headers) {
 208       rewind($fd);
 209     }
 210
 211     $sql   = NULL;
 212     $first = TRUE;
 213     $count = 0;
 214     while ($row = fgetcsv($fd, 0, $fieldSeparator)) {
 215       // skip rows that dont match column count, else we get a sql error
 216       if (count($row) != $numColumns) {
 217         continue;
 218       }
 219
 220       if (!$first) {
 221         $sql .= ', ';
 222       }
 223
 224       $first = FALSE;
 225       $row = array_map('civicrm_mysql_real_escape_string', $row);
 226       $sql .= "('" . implode("', '", $row) . "')";
 227       $count++;
 228
 229       if ($count >= self::NUM_ROWS_TO_INSERT && !empty($sql)) {
 230         $sql = "INSERT IGNORE INTO $table VALUES $sql";
 231         $db->query($sql);
 232
 233         $sql   = NULL;
 234         $first = TRUE;
 235         $count = 0;
 236       }
 237     }
 238
 239     if (!empty($sql)) {
 240       $sql = "INSERT IGNORE INTO $table VALUES $sql";
 241       $db->query($sql);
 242     }
 243
 244     fclose($fd);
 245
 246     //get the import tmp table name.
 247     $result['import_table_name'] = $table;
 248
 249     return $result;
 250   }
 251 }
 252
 253 /**
 254  * @param $string
 255  *
 256  * @return string
 257  */
 258 function civicrm_mysql_real_escape_string($string) {
 259   static $dao = NULL;
 260   if (!$dao) {
 261     $dao = new CRM_Core_DAO();
 262   }
 263   return $dao->escape($string);
 264 }