includes/utils/BatchRowIterator.php

   1 <?php
   2 /**
   3  * Allows iterating a large number of rows in batches transparently.
   4  * By default when iterated over returns the full query result as an
   5  * array of rows.  Can be wrapped in RecursiveIteratorIterator to
   6  * collapse those arrays into a single stream of rows queried in batches.
   7  *
   8  * This program is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * This program is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with this program; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  21  * http://www.gnu.org/copyleft/gpl.html
  22  *
  23  * @file
  24  * @ingroup Maintenance
  25  */
  26 class BatchRowIterator implements RecursiveIterator {
  27
  28         /**
  29          * @var IDatabase $db The database to read from
  30          */
  31         protected $db;
  32
  33         /**
  34          * @var string $table The name of the table to read from
  35          */
  36         protected $table;
  37
  38         /**
  39          * @var array $primaryKey The name of the primary key(s)
  40          */
  41         protected $primaryKey;
  42
  43         /**
  44          * @var integer $batchSize The number of rows to fetch per iteration
  45          */
  46         protected $batchSize;
  47
  48         /**
  49          * @var array $conditions Array of strings containing SQL conditions
  50          *  to add to the query
  51          */
  52         protected $conditions = [];
  53
  54         /**
  55          * @var array $joinConditions
  56          */
  57         protected $joinConditions = [];
  58
  59         /**
  60          * @var array $fetchColumns List of column names to select from the
  61          *  table suitable for use with IDatabase::select()
  62          */
  63         protected $fetchColumns;
  64
  65         /**
  66          * @var string $orderBy SQL Order by condition generated from $this->primaryKey
  67          */
  68         protected $orderBy;
  69
  70         /**
  71          * @var array $current The current iterator value
  72          */
  73         private $current = [];
  74
  75         /**
  76          * @var integer key 0-indexed number of pages fetched since self::reset()
  77          */
  78         private $key;
  79
  80         /**
  81          * @param IDatabase $db The database to read from
  82          * @param string       $table      The name of the table to read from
  83          * @param string|array $primaryKey The name or names of the primary key columns
  84          * @param integer      $batchSize  The number of rows to fetch per iteration
  85          * @throws MWException
  86          */
  87         public function __construct( IDatabase $db, $table, $primaryKey, $batchSize ) {
  88                 if ( $batchSize < 1 ) {
  89                         throw new MWException( 'Batch size must be at least 1 row.' );
  90                 }
  91                 $this->db = $db;
  92                 $this->table = $table;
  93                 $this->primaryKey = (array)$primaryKey;
  94                 $this->fetchColumns = $this->primaryKey;
  95                 $this->orderBy = implode( ' ASC,', $this->primaryKey ) . ' ASC';
  96                 $this->batchSize = $batchSize;
  97         }
  98
  99         /**
 100          * @param array $condition Query conditions suitable for use with
 101          *  IDatabase::select
 102          */
 103         public function addConditions( array $conditions ) {
 104                 $this->conditions = array_merge( $this->conditions, $conditions );
 105         }
 106
 107         /**
 108          * @param array $condition Query join conditions suitable for use
 109          *  with IDatabase::select
 110          */
 111         public function addJoinConditions( array $conditions ) {
 112                 $this->joinConditions = array_merge( $this->joinConditions, $conditions );
 113         }
 114
 115         /**
 116          * @param array $columns List of column names to select from the
 117          *  table suitable for use with IDatabase::select()
 118          */
 119         public function setFetchColumns( array $columns ) {
 120                 // If it's not the all column selector merge in the primary keys we need
 121                 if ( count( $columns ) === 1 && reset( $columns ) === '*' ) {
 122                         $this->fetchColumns = $columns;
 123                 } else {
 124                         $this->fetchColumns = array_unique( array_merge(
 125                                 $this->primaryKey,
 126                                 $columns
 127                         ) );
 128                 }
 129         }
 130
 131         /**
 132          * Extracts the primary key(s) from a database row.
 133          *
 134          * @param stdClass $row An individual database row from this iterator
 135          * @return array Map of primary key column to value within the row
 136          */
 137         public function extractPrimaryKeys( $row ) {
 138                 $pk = [];
 139                 foreach ( $this->primaryKey as $column ) {
 140                         $pk[$column] = $row->$column;
 141                 }
 142                 return $pk;
 143         }
 144
 145         /**
 146          * @return array The most recently fetched set of rows from the database
 147          */
 148         public function current() {
 149                 return $this->current;
 150         }
 151
 152         /**
 153          * @return integer 0-indexed count of the page number fetched
 154          */
 155         public function key() {
 156                 return $this->key;
 157         }
 158
 159         /**
 160          * Reset the iterator to the begining of the table.
 161          */
 162         public function rewind() {
 163                 $this->key = -1; // self::next() will turn this into 0
 164                 $this->current = [];
 165                 $this->next();
 166         }
 167
 168         /**
 169          * @return boolean True when the iterator is in a valid state
 170          */
 171         public function valid() {
 172                 return (bool)$this->current;
 173         }
 174
 175         /**
 176          * @return boolean True when this result set has rows
 177          */
 178         public function hasChildren() {
 179                 return $this->current && count( $this->current );
 180         }
 181
 182         /**
 183          * @return RecursiveIterator
 184          */
 185         public function getChildren() {
 186                 return new NotRecursiveIterator( new ArrayIterator( $this->current ) );
 187         }
 188
 189         /**
 190          * Fetch the next set of rows from the database.
 191          */
 192         public function next() {
 193                 $res = $this->db->select(
 194                         $this->table,
 195                         $this->fetchColumns,
 196                         $this->buildConditions(),
 197                         __METHOD__,
 198                         [
 199                                 'LIMIT' => $this->batchSize,
 200                                 'ORDER BY' => $this->orderBy,
 201                         ],
 202                         $this->joinConditions
 203                 );
 204
 205                 // The iterator is converted to an array because in addition to
 206                 // returning it in self::current() we need to use the end value
 207                 // in self::buildConditions()
 208                 $this->current = iterator_to_array( $res );
 209                 $this->key++;
 210         }
 211
 212         /**
 213          * Uses the primary key list and the maximal result row from the
 214          * previous iteration to build an SQL condition sufficient for
 215          * selecting the next page of results.  All except the final key use
 216          * `=` conditions while the final key uses a `>` condition
 217          *
 218          * Example output:
 219          *        array( '( foo = 42 AND bar > 7 ) OR ( foo > 42 )' )
 220          *
 221          * @return array The SQL conditions necessary to select the next set
 222          *  of rows in the batched query
 223          */
 224         protected function buildConditions() {
 225                 if ( !$this->current ) {
 226                         return $this->conditions;
 227                 }
 228
 229                 $maxRow = end( $this->current );
 230                 $maximumValues = [];
 231                 foreach ( $this->primaryKey as $column ) {
 232                         $maximumValues[$column] = $this->db->addQuotes( $maxRow->$column );
 233                 }
 234
 235                 $pkConditions = [];
 236                 // For example: If we have 3 primary keys
 237                 // first run through will generate
 238                 //   col1 = 4 AND col2 = 7 AND col3 > 1
 239                 // second run through will generate
 240                 //   col1 = 4 AND col2 > 7
 241                 // and the final run through will generate
 242                 //   col1 > 4
 243                 while ( $maximumValues ) {
 244                         $pkConditions[] = $this->buildGreaterThanCondition( $maximumValues );
 245                         array_pop( $maximumValues );
 246                 }
 247
 248                 $conditions = $this->conditions;
 249                 $conditions[] = sprintf( '( %s )', implode( ' ) OR ( ', $pkConditions ) );
 250
 251                 return $conditions;
 252         }
 253
 254         /**
 255          * Given an array of column names and their maximum value  generate
 256          * an SQL condition where all keys except the last match $quotedMaximumValues
 257          * exactly and the last column is greater than the matching value in
 258          * $quotedMaximumValues
 259          *
 260          * @param array $quotedMaximumValues The maximum values quoted with
 261          *  $this->db->addQuotes()
 262          * @return string An SQL condition that will select rows where all
 263          *  columns match the maximum value exactly except the last column
 264          *  which must be greater than the provided maximum value
 265          */
 266         protected function buildGreaterThanCondition( array $quotedMaximumValues ) {
 267                 $keys = array_keys( $quotedMaximumValues );
 268                 $lastColumn = end( $keys );
 269                 $lastValue = array_pop( $quotedMaximumValues );
 270                 $conditions = [];
 271                 foreach ( $quotedMaximumValues as $column => $value ) {
 272                         $conditions[] = "$column = $value";
 273                 }
 274                 $conditions[] = "$lastColumn > $lastValue";
 275
 276                 return implode( ' AND ', $conditions );
 277         }
 278 }