MySQL 函数
在线手册：中文 英文
PHP手册

mysql_affected_rows

(PHP 4, PHP 5)

mysql_affected_rows — 取得前一次 MySQL 操作所影响的记录行数

说明

int mysql_affected_rows ([ resource $link_identifier ] )

取得最近一次与 link_identifier 关联的 INSERT，UPDATE 或 DELETE 查询所影响的记录行数。

参数

link_identifier

MySQL 连接。如不指定连接标识，则使用由 mysql_connect() 最近打开的连接。如果没有找到该连接，会尝试不带参数调用 mysql_connect() 来创建。如没有找到连接或无法建立连接，则会生成 E_WARNING 级别的错误。

返回值

执行成功则返回受影响的行的数目，如果最近一次查询失败的话，函数返回 -1。

如果最近一次操作是没有任何条件（WHERE）的 DELETE 查询，在表中所有的记录都会被删除，但本函数返回值在 4.1.2 版之前都为 0。

当使用 UPDATE 查询，MySQL 不会将原值和新值一样的列更新。这样使得 mysql_affected_rows() 函数返回值不一定就是查询条件所符合的记录数，只有真正被修改的记录数才会被返回。

REPLACE 语句首先删除具有相同主键的记录，然后插入一个新记录。本函数返回的是被删除的记录数加上被插入的记录数。

范例

Example #1 mysql_affected_rows() 例子

<?php
$link = mysql_connect('localhost', 'mysql_user', 'mysql_password');
if (!$link) {
    die('Could not connect: ' . mysql_error());
}
mysql_select_db('mydb');

/* 本例返回被删除记录的准确数目 */
mysql_query('DELETE FROM mytable WHERE id < 10');
printf("Records deleted: %d\n", mysql_affected_rows());

/* 对于非真值的 WHERE 子句，应返回 0 */
mysql_query('DELETE FROM mytable WHERE 0');
printf("Records deleted: %d\n", mysql_affected_rows());
?>

以上例程的输出类似于：

Records deleted: 10
Records deleted: 0

Example #2 使用事务处理的 mysql_affected_rows() 例子

<?php
$link = mysql_connect('localhost', 'mysql_user', 'mysql_password');
if (!$link) {
    die('Could not connect: ' . mysql_error());
}
mysql_select_db('mydb');

/* Update records */
mysql_query("UPDATE mytable SET used=1 WHERE id < 10");
printf ("Updated records: %d\n", mysql_affected_rows());
mysql_query("COMMIT");
?>

以上例程的输出类似于：

Updated Records: 10

注释

Note: Suggested alternatives

Use of this extension is discouraged. Instead, the MySQLi or PDO_MySQL extension should be used. See also MySQL: choosing an API for more information.

Alternatives to this function include:

Note: 事务处理

如果使用事务处理（transactions），需要在 INSERT，UPDATE 或 DELETE 查询后调用 mysql_affected_rows() 函数，而不是在 COMMIT 命令之后。

Note: SELECT 语句

要获取 SELECT 所返回的行数，可以用 mysql_num_rows()。

参见

MySQL 函数
在线手册：中文 英文
PHP手册
PHP手册 - N: 取得前一次 MySQL 操作所影响的记录行数

用户评论:

cbal at freemail dot hu (07-Feb-2012 09:21)

I dont know why but sometimes mysql_affected_rows does not work (return 0 or nothing) besides the query worked well.
So use this if you have a problem:

$result = mysql_query("update/delete ...");
$last = mysql_query("SELECT ROW_COUNT();");
$last = mysql_fetch_array($last);

Ome Ko (28-Jun-2011 09:10)

There are no rows affected by an update with identical data.
So here is one very ugly solution for these cases:
<?
function mysql_matched_rows() {
   $_kaBoom=explode(' ',mysql_info());
   return $_kaBoom[2];
}
?>

info at fedushin dot ru (13-Feb-2010 02:20)

mysql_affected_rows() DOES NOT count rows affected implicitly through 'ON DELETE CASCADE' and/or 'ON UPDATE CASCADE' foreign keys.
For example:

CREATE TABLE `types` (
  `type` varchar(10) NOT NULL,
  PRIMARY KEY (`type`)
) ENGINE=InnoDB;

CREATE TABLE `symbols` (
  `symbol` char(1) NOT NULL,
  `type` varchar(10) NOT NULL,
  PRIMARY KEY (`symbol`),
  KEY `FK_symbol_type` (`type`)
) ENGINE=InnoDB;

ALTER TABLE `symbols`
  ADD CONSTRAINT `FK_symbol_type` FOREIGN KEY (`type`) REFERENCES `types` (`type`) ON DELETE CASCADE ON UPDATE CASCADE;

INSERT INTO `types` VALUES ('Number'), ('Letter');
INSERT INTO `symbols` VALUES ('1', 'Number'), ('2', 'Number'), ('A', 'Letter'), ('B', 'Letter');

<?php
mysql_query('UPDATE types SET type = "Digit" WHERE type = "Number"');
echo mysql_affected_rows() . '<br>';

mysql_query('DELETE FROM types WHERE type = "Letter"');
echo mysql_affected_rows() . '<br>';
?>

Each query actually affects 3 rows (= 1 type + 2 symbols), but output is:
1
1

PaulD (08-Nov-2009 04:35)

Just a note: [you should] check for a mysql_affected_rows() return value of -1. If a series of queries were run as a transaction and one query failed, the logic in the foreach loop of the transaction() method in class MySQLDB will not issue a ROLLBACK. It may also be slightly more efficient to stop processing queries on the first failure.

sean at adtools dot co dot uk (09-Sep-2008 09:48)

Here's a little function I've been using for a while now, pass it two parameters (action command (1 or 0 see notes)) and a sql statement.

It returns a simple line which shows the length of time taken to action the query, the status of the query (0= query not actioned, you can set this value for testing, 1=success qry executed successfully, -1= failed, there was a problem with the sql statement) the number of lines affected by that query and the sql statement itself.

I've found this invaluable when trying to tie down large amounts of updates to a table, using this you can easily see where a query was successfully executed and the number of rows are affected, or where there are problems and a statement has failed for example.

<?php
function dosql($action,$sql){
  # assuming you have setup a link to your database entitled $link
  # action = 1 run this query
  # action = 0 don't run, just return sql statement
 
  $start = getmtime();
 
  if($action==1){
    $result = mysql_query($sql);
    $affectedrows = "[".mysql_affected_rows($link)."]";
  }
  return "[".number_format((getmtime()-$start),3)."][$action]: $sql\n";
  mysql_free_result($result);
}
?>

Example output:
[0.072][1][80]: UPDATE MYTABLE SET FIELD = 1;
[0.106][1][758]: UPDATE ANOTHERTABLE SET FIELD = 2;
[0.006][-1][0]: UPDATER ANOTHERTABLE SET FIELD = 2;

The output shows:

[Timetaken][result]][lines affected]

The result will be either -1, 0 or 1, -1 means there's a problem with the sql statement, 1 means it executed correctly, 0 means it wasn't executed.

EToS (09-Aug-2007 09:57)

i found a pretty nice way, this db class/function will accept an array of arrays of querys, it will auto check every line for affected rows in db, if one is 0 it will rollback and return false, else it will commit and return true, the call to the function is simple and is easy to read etc
----------

<?php
class MySQLDB
{
   private $connection;          // The MySQL database connection

   /* Class constructor */
   function MySQLDB(){
      /* Make connection to database */
      $this->connection = mysql_connect(DB_SERVER, DB_USER, DB_PASS) or die(mysql_error());
      mysql_select_db(DB_NAME, $this->connection) or die(mysql_error());
   }

   /* Transactions functions */

   function begin(){
         $null = mysql_query("START TRANSACTION", $this->connection);
      return mysql_query("BEGIN", $this->connection);
   }

   function commit(){
      return mysql_query("COMMIT", $this->connection);
   }
  
   function rollback(){
      return mysql_query("ROLLBACK", $this->connection);
   }

   function transaction($q_array){
         $retval = 1;

      $this->begin();

         foreach($q_array as $qa){
            $result = mysql_query($qa['query'], $this->connection);
            if(mysql_affected_rows() == 0){ $retval = 0; }
         }

      if($retval == 0){
         $this->rollback();
         return false;
      }else{
         $this->commit();
         return true;
      }
   }

};

/* Create database connection object */
$database = new MySQLDB;

// then from anywhere else simply put the transaction queries in an array or arrays like this:

   function function(){
      global $database;

      $q = array (
         array("query" => "UPDATE table WHERE something = 'something'"),
         array("query" => "UPDATE table WHERE something_else = 'something_else'"),
         array("query" => "DELETE FROM table WHERE something_else2 = 'something_else2'"),
      );

      $database->transaction($q);

   }
?>

HMax (02-Jul-2007 11:21)

If you use "INSERT INTO ... ON DUPLICATE KEY UPDATE" syntax, mysql_affected_rows() will return you 2 if the UPDATE was made (just as it does with the "REPLACE INTO" syntax) and 1 if the INSERT was.

So if you use one SQL request to insert several rows at a time, and some are inserted, some are just updated, you won't get the real count.

dobrys at abv dot bg (28-May-2007 10:35)

I see that when try to use mysql_affected_rows() with "mysql_pconnect(...)" without link indetifier as param in "mysql_affected_rows()" the result is allways -1.
When use link identifier "mysql_affected_rows($this_sql_connection)" - everything is Fine. This is is on PHP Version 5.2.0
Hope that this was helpfull for somebody

Typer85 at gmail dot com (31-Dec-2006 04:35)

Just to clarify about the possible return values in this Manual those not familiar with PHP and MySQL.

"-1 indicates that the query returned an error."

-1 will be returned if the query itself can not be issued to the server, possibly because of syntax error AND if the last query was not either an Insert or Update statement.

mlugassy at 2find dot co dot il (08-Oct-2005 02:22)

To solve the affectedRows() issue on MySQL using PEAR::DB, simply add a 'client_flags' key with a value of 2 to your $dsn options:

<?php
$dsn = array(
    'phptype'  => 'mysql',
    'client_flags'  => 2,
    'username' => 'someuser',
    'password' => 'apasswd',
    'hostspec' => 'localhost',
    'database' => 'thedb',
);
?>

temp02 at flexis dot com dot br (29-Jun-2005 01:39)

SCENARIO
1. You're using MySQL 4.1x with foreign keys.
2. You have table t2 linked to table t1 by a CASCADE ON DELETE foreign key.
3. t2 has a UNIQUE key so that duplicate records are unacceptable.
3. You have a REPLACE query on t1 followed by an INSERT query on t2 and expect the second query to fail if there's an attempted insert of a duplicate record.

PROBLEM
You notice that the second query is not failing as you had expected even though the record being inserted is an exact duplicate of a record previously inserted.

CAUSE
When the first query (the REPLACE query) deletes a record from t1 in the first stage of the REPLACE operation, it cascades the delete to the record that would be duplicated in t2. The second query then does not fail because the "duplicate" record is no longer a duplicate, as the original one has just been deleted.

steffen at showsource dot dk (28-Sep-2004 11:20)

Using OPTIMIZE TABLE will also return true.
So, if you want to check the numbers of deleted records, use mysql_affected_rows() before OPTIMIZE TABLE

deponti A_T tiscalinet D0T it (07-Nov-2003 12:52)

It works also for REPLACE query,returning:
0 if the record it's already updated (0 record modified),
1 if the record it's new (1 record inserted),
2 if the record it's updated (2 operations: 1 deletion+ 1 insertion)

ben-xo at NOSPAMdubplatesNOSPAM dot org (21-Apr-2002 03:30)

mysql_affected_rows() reports on the number of rows affected by an in-place operation on the database, but mysql_num_rows() returns the number of rows in a MySQL record set (which is held by PHP after MySQL has generated it). This means that if you can do

<?php
$a = mysql_query("SELECT ...");
$b = mysql_query("SELECT ...");
if (mysql_unm_rows($a) > mysql_num_rows($b)) print "a is larger";
else print "b is larger";
?>

... but this does not make sense for the operations supported by mysql_affected_rows(), which reports on the status of the database connection as a whole.

Particularly note this:

<?php
$query = "UPDATE ...";
mysql_query($query);
print mysql_affected_rows(); // more than 0
mysql_query($query); // same query twice
print mysql_affected_rows(); // 0.
?>

.. this is because the 2nd time you execute the identical query, all the rows are already updated so no rows are affected the 2nd time.

I hope this clears up why mysql_num_rows() and mysql_affected_rows() are fundamentally different

dfylstra at frontsys dot com (13-Aug-2001 08:06)

mysql_affected_rows() also reports the number of rows changed by the LOAD DATA command.  If you use the IGNORE option in LOAD DATA and you know the number of rows in the input file, you can use mysql_affected_rows() to determine the number of rows that were ignored.