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];
}
?>
(PHP 4, PHP 5)
mysql_affected_rows — Retourne le nombre de lignes affectées lors de la dernière opération MySQL
Cette extension était obsolète en PHP 5.5.0, et a été supprimée en PHP 7.0.0. À la place, vous pouvez utiliser l'extension MySQLi ou l'extension PDO_MySQL. Voir aussi MySQL : choisir une API du guide. Alternatives à cette fonction :
Retourne le nombre de lignes affectées lors de la dernière
requête INSERT, UPDATE, REPLACE ou DELETE avec
link_identifier
.
link_identifier
La connexion MySQL.
S'il n'est pas spécifié, la dernière connexion ouverte avec la fonction
mysql_connect() sera utilisée. Si une telle connexion
n'est pas trouvée, la fonction tentera d'ouvrir une connexion, comme
si la fonction mysql_connect() avait été appelée sans argument.
Si aucune connexion n'est trouvée ou établie, une alerte de niveau
E_WARNING
sera générée.
Retourne le nombre de lignes affectées en cas de succès et -1 si la dernière requête a échoué.
Si la dernière requête était un DELETE sans clause WHERE, tous les enregistrements ont été effacés, mais cette fonction va retourner 0 avec les versions de MySQL antérieures à 4.1.2.
Lorsque vous utilisez UPDATE, MySQL ne mettra pas à jour les colonnes si la nouvelle valeur est identique à l'ancienne. Il est donc possible que mysql_affected_rows() ne représente pas forcément le nombre de lignes correspondantes mais plutôt le nombre de lignes qui ont été réellement affectées par la requête.
La requête REPLACE commence par effacer les enregistrements possédant la même clé primaire et ensuite, insert les nouveaux enregistrements. Cette fonction retourne le nombre d'enregistrements effacés ainsi que le nombre d'enregistrements insérés.
Dans le cas d'une requête "INSERT ... ON DUPLICATE KEY UPDATE", la valeur
retournée sera 1
si une insertion a été effectuée,
ou 2
si une mise à jour d'une ligne existante a été réalisée.
Exemple #1 Exemple avec mysql_affected_rows()
<?php
$link = mysql_connect('localhost', 'mysql_user', 'mysql_password');
if (!$link) {
die('Impossible de se connecter : ' . mysql_error());
}
mysql_select_db('mydb');
/* Ceci devrait retourner le nombre correct de lignes effacées */
mysql_query('DELETE FROM mytable WHERE id < 10');
printf("Lignes effacées : %d\n", mysql_affected_rows());
/* avec la clause where qui n'est jamais vraie, on devrait obtenir 0 */
mysql_query('DELETE FROM mytable WHERE 0');
printf("Lignes effacées : %d\n", mysql_affected_rows());
?>
Résultat de l'exemple ci-dessus est similaire à :
Lignes effacées : 10 Lignes effacées : 0
Exemple #2 Exemple avec mysql_affected_rows() en utilisant les transactions
<?php
$link = mysql_connect('localhost', 'mysql_user', 'mysql_password');
if (!$link) {
die('Impossible de se connecter : ' . mysql_error());
}
mysql_select_db('mydb');
/* Mise à jour des lignes */
mysql_query("UPDATE mytable SET used=1 WHERE id < 10");
printf ("Lignes modifiées : %d\n", mysql_affected_rows());
mysql_query("COMMIT");
?>
Résultat de l'exemple ci-dessus est similaire à :
Lignes modifiées : 10
Note: Transactions
Si vous utilisez des transactions, vous devez appeler mysql_affected_rows() après votre requête INSERT, UPDATE ou DELETE et non après le COMMIT.
Note: Requêtes SELECT
Pour obtenir le nombre de lignes retournées par un SELECT, il est possible d'utiliser mysql_num_rows().
Note: Clés étrangères en cascade
mysql_affected_rows() ne compte pas les enregistrements affectés implicitement par un ON DELETE CASCADE et/ou un ON UPDATE CASCADE, concernant les clés étrangères.
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 $_kaBoom[2];
}
?>
SCENARIO
1. You're using MySQL 4.1x with foreign keys.
2. You have table t2 linked to table t1 by a CASCADE ON DELETE foreign key.
3. t2 has a UNIQUE key so that duplicate records are unacceptable.
3. You have a REPLACE query on t1 followed by an INSERT query on t2 and expect the second query to fail if there's an attempted insert of a duplicate record.
PROBLEM
You notice that the second query is not failing as you had expected even though the record being inserted is an exact duplicate of a record previously inserted.
CAUSE
When the first query (the REPLACE query) deletes a record from t1 in the first stage of the REPLACE operation, it cascades the delete to the record that would be duplicated in t2. The second query then does not fail because the "duplicate" record is no longer a duplicate, as the original one has just been deleted.
I was just testing "INSERT INTO ... ON DUPLICATE KEY UPDATE" syntax, on PHP 5.3.29 and mysql_affected_rows() was returning either 2 for updated row, 1 for inserted new row, and also 0, which was not documented, evidently when nothing was inserted. I was inserting a single row.
If you use "INSERT INTO ... ON DUPLICATE KEY UPDATE" syntax, mysql_affected_rows() will return you 2 if the UPDATE was made (just as it does with the "REPLACE INTO" syntax) and 1 if the INSERT was.
So if you use one SQL request to insert several rows at a time, and some are inserted, some are just updated, you won't get the real count.
I see that when try to use mysql_affected_rows() with "mysql_pconnect(...)" without link indetifier as param in "mysql_affected_rows()" the result is allways -1.
When use link identifier "mysql_affected_rows($this_sql_connection)" - everything is Fine. This is is on PHP Version 5.2.0
Hope that this was helpfull for somebody
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)
Note that when the CLIENT_FOUND_ROWS connection flag was used, affected_rows returns the number of rows matched by the WHERE condition of an UPDATE query, even if the query doesn't actually change those rows. I.e. for
INSERT INTO t(id, val) VALUES (1, 'x');
UPDATE t SET val = 'x' WHERE id = 1;
the number of affected rows will be 0 normally but 1 with CLIENT_FOUND_ROWS.
calling mysql_affected_rows(null)
is not the same that calling mysql_affected_rows()
So, if you have a $link variable that could be null, you must write
if($link)
$n=mysql_affected_rows($link);
else
$n=mysql_affected_rows();
In the case of INSERT where a row/slot had been previously deleted, making an uncollapsed hole in the table, and the record being inserted fills that empty row/slot, that is to say, the inserted data did not create a new row/slot/space, then this may explain why a zero result is returned by this function.
Using OPTIMIZE TABLE will also return true.
So, if you want to check the numbers of deleted records, use mysql_affected_rows() before OPTIMIZE TABLE
Here's a little function I've been using for a while now, pass it two parameters (action command (1 or 0 see notes)) and a sql statement.
It returns a simple line which shows the length of time taken to action the query, the status of the query (0= query not actioned, you can set this value for testing, 1=success qry executed successfully, -1= failed, there was a problem with the sql statement) the number of lines affected by that query and the sql statement itself.
I've found this invaluable when trying to tie down large amounts of updates to a table, using this you can easily see where a query was successfully executed and the number of rows are affected, or where there are problems and a statement has failed for example.
<?php
function dosql($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.