©
このドキュメントでは、 php中国語ネットマニュアル リリース
(PHP 5 >= 5.1.0, PECL pdo >= 0.1.0)
PDOStatement::rowCount — 返回受上一个 SQL 语句影响的行数
PDOStatement::rowCount() 返回上一个由对应的 PDOStatement 对象执行DELETE、 INSERT、或 UPDATE 语句受影响的行数。
如果上一条由相关 PDOStatement 执行的 SQL 语句是一条 SELECT 语句,有些数据可能返回由此语句返回的行数。但这种方式不能保证对所有数据有效,且对于可移植的应用不应依赖于此方式。
返回行数。
Example #1 返回删除的行数
PDOStatement::rowCount() 返回受 DELETE、INSERT、 或 UPDATE 语句影响的行数。
<?php
$del = $dbh -> prepare ( 'DELETE FROM fruit' );
$del -> execute ();
print( "Return number of rows that were deleted:\n" );
$count = $del -> rowCount ();
print( "Deleted $count rows.\n" );
?>
以上例程会输出:
Return number of rows that were deleted: Deleted 9 rows.
Example #2 计算由一个 SELECT 语句返回的行数
对于大多数数据库, PDOStatement::rowCount() 不能返回受一条 SELECT 语句影响的行数。替代的方法是,使用 PDO::query() 来发出一条和原打算中的SELECT语句有相同条件表达式的 SELECT COUNT(*) 语句,然后用 PDOStatement::fetchColumn() 来取得返回的行数。这样应用程序才能正确执行。
<?php
$sql = "SELECT COUNT(*) FROM fruit WHERE calories > 100" ;
if ( $res = $conn -> query ( $sql )) {
if ( $res -> fetchColumn () > 0 ) {
$sql = "SELECT name FROM fruit WHERE calories > 100" ;
foreach ( $conn -> query ( $sql ) as $row ) {
print "Name: " . $row [ 'NAME' ] . "\n" ;
}
}
else {
print "No rows matched the query." ;
}
}
$res = null ;
$conn = null ;
?>
以上例程会输出:
apple banana orange pear
[#1] @rhavendc (twitter) [2015-11-23 08:10:56]
We're having problem with these PDOStatement::fetchColumn() and PDOStatement::rowCount(). I don't know if we have alike case to others or this is just a problem in our coding. In local, the rowCount() is not giving the correct number of rows but it is working fine when uploaded to our hosting site.. while the fetchColumn() is the reverse. The fetchColumn() is working fine in local but not anymore when uploaded. I don't know what is really going on with that but I think rowCount() is the best and the others are optional to use.
[#2] info at buylikesandviews dot com [2015-03-21 09:04:13]
every good work
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.
[#3] phpnet at maya-control dot ro [2014-04-10 11:09:20]
Yet another workaround to return the row count inside only ONE select (see limitations below!):
$sth = $dbh->prepare("SELECT *,count(*) AS howmany FROM users WHERE email=:email and password=:pass"); #var placeholders
$sth->execute(array(':email'=>$email, ':pass'=>$pass)); #var binding
$row = $sth->fetch(); #get one row (it'll always be one and only one!!!)
if ($row['howmany'] == 1){ #we have a match and only one! cool!
echo $row['email'], $row['name'], $row['phone'], ... ;
} elseif ($row['howmany']>1) { #more than one row returned
#one programmer should be fired 'cause he's not checking for
#for existing emails, before creating a new user
...
# treat this exception somehow or simply skip this branch,
# if you're sure it won't happen in your table
} else { #no match in the table ($row['howmany'] == 0)
echo "Email/pass didn't match the ones in the database!";
}
Advantages:
- only one select statement is executed, no two steps needed!
- it checks if one row exists in the table or not, according to the WHERE clause.
- it returns all (or only a selection of) fields for that one row, if exists.
Disadvantages:
- it doesn't return row fields reliable if more than one row found. If more than one row responds to the SELECT query, the query returns still only one row and you don't know which one exactly .
Maybe using a SORT BY, would make it a bit more predictible (as in: "if more than one users found, return the last user added in the table") but it's more a matter of good design of the program that fills in the table initially.
Uses:
- It is perfect for checking if a user/pass pair is present in a users table and to return the other fields of the user (like name, phone, whatever) if user was found.
[#4] sERGE-01 [2013-11-04 12:15:19]
In some drivers rowCount() only works when using the prepare() with PDO::CURSOR_SCROLL
So, you can modify PDO class:
<?php
class myPDO extends PDO
{
function query($query, $values=null)
{
if($query == "")
return false;
if($sth = $this->prepare($query, array(PDO::ATTR_CURSOR => PDO::CURSOR_SCROLL)))
{
$res = ($values) ? $sth->execute($values) : $sth->execute();
if(!$res)
return false;
}
return $sth;
}
}
?>
Now let's test (i using php 5.2.9-2):
<?php
function TestRowCount($dsn, $db_user, $db_pass)
{
$pdh = new PDO($dsn, $db_user, $db_pass);
$sth = $pdh->query("SELECT * FROM sys.tables");
print "rowCount() Standart: ".$sth->rowCount()."<br>";
$pdh = new myPDO($dsn, $db_user, $db_pass);
$sth = $pdh->query("SELECT * FROM sys.tables");
print "rowCount() New: ".$sth->rowCount()."<br><br>";
$pdh=null;
}
$db_server = "xxx";
$db_name = "xxx";
$db_user = "xxx";
$db_pass = "xxx";
print "PDO_MSSQL"."<br>";
TestRowCount("mssql:host=$db_server;dbname=$db_name", $db_user, $db_pass);
print "MSSQL throw PDO_ODBC"."<br>";
TestRowCount("odbc:DRIVER={SQL Server};SERVER=$db_server;DATABASE=$db_name;", $db_user, $db_pass);
print "MS SQL driver 2.0"."<br>";
TestRowCount("sqlsrv:server=$db_server;Database=$db_name", $db_user, $db_pass);
?>
My results:
-------------------
PDO_MSSQL
rowCount() Standart: 0
rowCount() New: 0
MSSQL throw PDO_ODBC
rowCount() Standart: -1
rowCount() New: 53
MS SQL driver 2.0
rowCount() Standart: -1
rowCount() New: 53
-------------------
With myPDO class you can use prepared queries like:
<?php
$pdh = new myPDO($dsn, $db_user, $db_pass);
$sth = $pdh->query("select * from data where id>? or name like ?", array(100, "A%"));
?>
[#5] user at nospam dot example dot com [2013-09-03 18:49:57]
MySQL does not seem to return anything in rowCount for a select statement, but you can easily and efficiently get the row count as follows:
class db extends PDO {
public function last_row_count() {
return $this->query("SELECT FOUND_ROWS()")->fetchColumn();
}
}
$myDb = new db('mysql:host=myhost;dbname=mydb', 'login', 'password' );
Then, after running your query:
if ( $myDb->last_row_count() == 0 ) {
echo "Do something!";
}
[#6] Daniel Karp [2012-08-31 00:09:13]
Note that an INSERT ... ON DUPLICATE KEY UPDATE statement is not an INSERT statement, rowCount won't return the number or rows inserted or updated for such a statement. For MySQL, it will return 1 if the row is inserted, and 2 if it is updated, but that may not apply to other databases.
[#7] leandro at marquesini dot com [2012-01-03 10:04:01]
To display information only when the query is not empty, I do something like this:
<?php
$sql = 'SELECT model FROM cars';
$stmt = $db->prepare($sql);
$stmt->execute();
if ($data = $stmt->fetch()) {
do {
echo $data['model'] . '<br>';
} while ($data = $stmt->fetch());
} else {
echo 'Empty Query';
}
?>
[#8] Ome Ko [2011-07-16 12:08:41]
When updating a Mysql table with identical values nothing's really affected so rowCount will return 0. As Mr. Perl below noted this is not always preferred behaviour and you can change it yourself since PHP 5.3.
Just create your PDO object with
<?php php
$p = new PDO($dsn, $u, $p, array(PDO::MYSQL_ATTR_FOUND_ROWS => true));
?>
and rowCount() will tell you how many rows your update-query actually found/matched.
[#9] e dot sand at elisand dot com [2008-11-19 11:32:36]
My rowCount() workaround & how it's used:
<?php
class MyPDO extends PDO {
private $queryString;
public function query() {
$args = func_get_args();
$this->queryString = func_get_arg(0);
return call_user_func_array(array(&$this, 'parent::query'), $args);
}
public function rowCount() {
$regex = '/^SELECT\s+(?:ALL\s+|DISTINCT\s+)?(?:.*?)\s+FROM\s+(.*)$/i';
if (preg_match($regex, $this->queryString, $output) > 0) {
$stmt = parent::query("SELECT COUNT(*) FROM {$output[1]}", PDO::FETCH_NUM);
return $stmt->fetchColumn();
}
return false;
}
}
$pdo = new MyPDO("sqlite::memory:");
$result = $pdo->query("SELECT row1, row2 FROM table WHERE something = 5");
if ($pdo->rowCount() > 0) {
echo "{$result['row1']}, {$result['row2']}";
}
?>
[#10] e dot sand at elisand dot com [2008-11-19 11:30:27]
As of SQLite 3.x, the SQLite API itself changed and now all queries are implemented using "statements". Because of this, there is no way for PDO to know the rowCount of a SELECT result because the SQLite API itself doesn't offer this ability.
As a workaround, I created my own rowCount() function - it's a bit of a hack and hasn't been fully tested yet (I don't know how it will work when using JOINs in SELECTs, etc...), but at least alleviates the necessity for SELECT COUNT(*)'s everywhere in your code.
I would have preferred if it were possible to overload the rowCount() function from PDOStatement, but I don't think it's possible (or I don't know how to do it). There's also potential room for a bit more security ensuring that $queryString is wiped clean after other query()s so that you don't get a bad result, etc...
The actual code should be posted in the above/below post (max post limits, argh!). If others wish to extend/perfect this method, please keep me posted with an email as to what you've done.
[#11] gunnrosebutpeace at gmail dot com [2008-06-03 03:19:27]
It'd better to use SQL_CALC_FOUND_ROWS, if you only use MySQL. It has many advantages as you could retrieve only part of result set (via LIMIT) but still get the total row count.
code:
<?php
$db = new PDO(DSN...);
$db->setAttribute(array(PDO::MYSQL_USE_BUFFERED_QUERY=>TRUE));
$rs = $db->query('SELECT SQL_CALC_FOUND_ROWS * FROM table LIMIT 5,15');
$rs1 = $db->query('SELECT FOUND_ROWS()');
$rowCount = (int) $rs1->fetchColumn();
?>
[#12] Matt [2007-10-07 09:22:07]
Great, while using MySQL5, the only way to get the number of rows after doing a PDO SELECT query is to either execute a separate SELECT COUNT(*) query (or to do count($stmt->fetchAll()), which seems like a ridiculous waste of overhead and programming time.
Another gripe I have about PDO is its inability to get the value of output parameters from stored procedures in some DBMSs, such as SQL Server.
I'm not so sure I'm diggin' PDO yet.