mysql_insert_id

(PHP 3, PHP 4, PHP 5)

mysql_insert_id --  取得上一步 INSERT 操作产生的 ID

说明

int mysql_insert_id ( [resource link_identifier] )

mysql_insert_id() 返回给定的 link_identifier 中上一步 INSERT 查询中产生的 AUTO_INCREMENT 的 ID 号。如果没有指定 link_identifier,则使用上一个打开的连接。

如果上一查询没有产生 AUTO_INCREMENT 的值,则 mysql_insert_id() 返回 0。如果需要保存该值以后使用,要确保在产生了值的查询之后立即调用 mysql_insert_id()

注: MySQL 中的 SQL 函数 LAST_INSERT_ID() 总是保存着最新产生的 AUTO_INCREMENT 值,并且不会在查询语句之间被重置。

警告

mysql_insert_id() 将 MySQL 内部的 C API 函数 mysql_insert_id() 的返回值转换成 long(PHP 中命名为 int)。如果 AUTO_INCREMENT 的列的类型是 BIGINT,则 mysql_insert_id() 返回的值将不正确。可以在 SQL 查询中用 MySQL 内部的 SQL 函数 LAST_INSERT_ID() 来替代。

例子 1. mysql_insert_id() 例子

<?php
    mysql_connect
("localhost", "mysql_user", "mysql_password") or
        die(
"Could not connect: " . mysql_error());
    
mysql_select_db("mydb");

    
mysql_query("INSERT INTO mytable (product) values ('kossu')");
    
printf ("Last inserted record has id %d\n", mysql_insert_id());
?>

参见 mysql_query()


add a note add a note User Contributed Notes
buana95 att yahoo dott com
29-Oct-2006 05:57
It's not true that mysql_insert_id() only returns the ID generated for an AUTO_INCREMENT column by the previous INSERT query.

We can use LAST_INSERT_ID() statement that will return value for mysql_insert_id().

Example:

$sql = "UPDATE `mytable` SET `mytable_id`= LAST_INSERT_ID(`mytable_id` + 1) WHERE ...;

@mysql_query($sql);
$last_id = mysql_insert_id();

LAST_INSERT_ID() statement will affect mysql_insert_id().

Regards,
Buana
mariano at uvcms dot com
20-Sep-2006 12:01
When used in transactions, mysql_insert_id() MUST be called before committing. Otherwise, it will return unpredictable results.
squezy16 at hotmail dot com
17-Sep-2006 11:46
One of the problem of the LAST_INSERT_ID function is that it may return an old Auto_Increment number if the INSERT did not have the time to run properly while the mysql_insert_id always returns the good value. I tried to insert a row imidiatly followed by a select LAST_INSERT_ID and it kept on returning me the LAST_INSERT_ID before my insert statement.
rudolflai at gmail dot com
11-Sep-2006 05:59
Hi,

Below is a snippet for retriving the next insert id from a mysql table. It functions like the auto_increment option in phpMyAdmin. Hope you find it helpful. =]

<?php

function mysql_autoid($id,$table){
  
$query = 'SELECT MAX('.$id.') AS last_id FROM '.$table;
  
$result = mysql_query($query);
  
$result = mysql_fetch_array($result);
   return
$result[last_id]+1;
}

/// usage mysql_autoid('nID','news');

?>

Cheers,
Rudolf
james dot baxter at nospamplease dot com
09-May-2006 08:09
Have had many people coming to me with issues regarding last_insert_id and mysql_insert_id() etc

This is a very simple solution that caters for most. As you can see it gets the last auto inc from the field, in this case, ID.

mysql_query("LOCK TABLES art WRITE");

mysql_query("INSERT INTO table (....) VALUES (....)");

$result = mysql_query("SELECT MAX(ID) AS LAST_ID FROM table");

$result = mysql_fetch_array($result);

mysql_query("UNLOCK TABLES");

echo "Last ID#" . $result[LAST_ID];
john k.
01-Feb-2006 06:55
Be careful when using "insert ignore". If the unique index already exists, the record will not be added, but it WILL give you the id of the next auto_increment value even though it didn't create it.

<?php
$sql
= "insert ignore into sometable set num=10";
mysql_query($sql) or die();
echo
mysql_affected_rows()."<br>";
echo
mysql_insert_id()."<br><br>";

// same record, database is unique on 'num'
$sql = "insert ignore into sometable set num=10";
mysql_query($sql) or die();
echo
mysql_affected_rows()."<br>";
echo
mysql_insert_id()."<br><br>";
?>

would give:
1
116372

0
116373
heiligkind at yahoo dot de
14-Dec-2005 05:31
If you insert a data row by using the ON DUPLICATE KEY UPDATE clause in an INSERT-statement, the mysql_insert_id() function will return not the same results as if you directly use LAST_INSERT_ID() in MySQL.

See the following example:

<?
  
// insert a datarow, primary key is auto_increment
   // value is a unique key
  
$query = "INSERT INTO test (value) VALUES ('test')";
  
mysql_query( $query );

   echo
'LAST_INSERT_ID: ',
        
mysql_query( "SELECT LAST_INSERT_ID()" ),
        
'<br>mysql_insert_id: ',
        
mysql_insert_id();

?>

This will print:

LAST_INSERT_ID: 1
mysql_insert_id: 1

In this case the function returns the same as the MySQL-Statement.
But see the insert on an existing key:

<?
   $query
= "INSERT INTO test (value)
                 VALUES ('test')
                 ON DUPLICATE KEY UPDATE value = 'test2'"
;
  
mysql_query( $query );

   echo
'LAST_INSERT_ID: ',
        
mysql_query( "SELECT LAST_INSERT_ID()" ),
        
'<br>mysql_insert_id: ',
        
mysql_insert_id();

?>

This will print:

LAST_INSERT_ID: 2
mysql_insert_id: 1

By using the ON DUPLICATE KEY UPDATE clause, only the old datarow will be modified, if the INSERT statement causes a duplicate entry, but the LAST_INSERT_ID() function returns the next auto_increment value for the primary key, which is by the way not set as the next auto_increment value in the database.

The mysql_insert_id() function returns the primary key of the old (and changed) data row. For me this is the right operation method, because the LAST_INSERT_ID() function returns a value which is not referenced to a data row at all.

Greets from Munich.

heiligkind
Soroushl at gmail dot com
07-Dec-2005 01:53
As mentioned by frumler at the-beach dot no_spam dot net

the LAST_INSERT_ID works like a charm when inserting values into tables.

I'm not sure why anyone would need mysql_insert_id() when LAST_INSERT_ID is readily available.

example:

Say you have a table called "transaction" and a table called "accounts".  Obviously each account must be created using a transaction, so every time a record is created in the accounts table, a record must be created in the transaction table containing the same account_id(auto_increment) that was just created by mysql.

Here's a simple way to do this:

<?php
$account_query
="INSERT INTO accounts (account_id,type) VALUES (NULL,'saving')";
$transaction_query="INSERT INTO transactions(transaction_id,account_id) VALUES (NULL,LAST_INSERT_ID)";

$a_query=mysql_query($account_query);
$t_query=mysql_query($transaction_query);

?>
03-Dec-2005 02:48
Take care of setting an empty value for the AUTO_INCREMENT Field else you never get a value except zero returned from mysq_insert_id() ....

Ciao Ephraim
bholbrook at servillian dot com
02-Sep-2005 04:50
My apologies for the error below (that was modified out of a class) - as you cannot define a constant as an array.

replace the line:

<?

define
("ID",...);

?>

with

<?

$mysql_id
= mysql_query("...");

?>

$mysql_id is now an array in which the first element $mysql_id[0] holds the last inserted id.

Apologies if anyone struggled over that one... esp. the noobs.
bholbrook at servillian dot com
31-Aug-2005 04:45
An effective way to retrieve the last insert id without fail or having two queries inserted at the same time causing an unexpected number being returned.

<?

mysql_query
("LOCK TABLES apc_forms WRITE");
mysql_query("SET AUTOCOMMIT = 0");
mysql_query("INSERT INTO apc_forms (form_title, form_event_id, form_expirey) VALUES ('title',1,'2005-10-10')");
define('ID',mysql_query("SELECT LAST_INSERT_ID()"));
mysql_query("COMMIT");
mysql_query("UNLOCK TABLES");

?>
athies at gmail dot com
22-Jul-2005 05:36
Just a quick note. mysql_insert_id() does work with REPLACE.
ed at is-cool dot net
09-May-2005 01:25
Beware, mysql_insert_id() only returns the value of the last syntaxically correct insert statement.

If your code has a problem and the insert is not understood by the server then the value of the last working insert command is returned.

Put something else in place such as "select count( id ) from table" before and after the mysql_insert_id() call to ensure that a row was inserted.
relic at daimi dot au dot dk
20-Feb-2005 10:57
A bit more on return values:
mysql_insert_id() returns 0 if you haven't actually manipulated anything yet.

Also, it returns 0 even if the DB connection is lost[0] between inserting and calling mysql_insert_id() - so you can always count on getting an integer.

[0] By 'lost' I mean e.g. a DB crash. Actually closing the open link and then trying to communicate with the DB will of course result in an error.
brodseba AT brodseba DOT com
16-Feb-2005 04:37
It's possible to do the same with M$ Server.

function odbc_insert_id()
{
  $query = "SELECT @@IDENTITY AS ID;";
  $result = odbc_exec($this->m_rConnectionID, $query);
  $row = odbc_fetch_object($result);
  return $row->ID;
}
sander [ad] deltaserv [d0t] nl
11-Nov-2004 09:38
In reply to: sly at noiretblanc dot org:

Make sure that auto_increment has an capital A as the first letter, otherwise it WON'T work! So you have to spell it as Auto_increment... And then it works fine.
Baak
01-Oct-2004 02:04
I believe the "resource link" being referred to is not what is returned from mysql_query() but the $link returned from mysql_connect(). mysql_insert_id() will just use the most recent connection if there is no explicit $link being used.

So the above example in the manual page itself should behave the same with mysql_insert_id($link) at the end instead of the mysql_insert_id() they used. If you had multiple connections, the $link might come in handy.

Also in reading the mysql manual itself, there is some enlightening information on the fact that this does appear to be totally safe to use because it is on a per-connection basis.

Here's the relevant quote from the manual on LAST_INSERT_ID() which is located here: http://dev.mysql.com/doc/mysql/en/Information_functions.html

"The last ID that was generated is maintained in the server on a per-connection basis. This means the value the function returns to a given client is the most recent AUTO_INCREMENT value generated by that client. The value cannot be affected by other clients, even if they generate AUTO_INCREMENT values of their own. This behavior ensures that you can retrieve your own ID without concern for the activity of other clients, and without the need for locks or transactions."

Sounds safe to me. I couldn't imagine this would be done any other way *but* on a per-connection basis, otherwise chaos would ensue. The only way to test it would be to perform a multi-thread type test. Perhaps someone is up for it and wants to post their results somewhere? :)
Steve Bond
25-Jun-2004 02:47
If you use this function after doing an INSERT ... SELECT to insert multiple rows at once, you get the autonumber ID of the *first* row added by the INSERT.

e.g. if there are 4 records in table 'init' that have column 'type' = 2
I want to add these 4 records to table 'game'
Table game has an autonumber column 'game_id' that is currently at 32.

If I do this query:

INSERT INTO game (type, players, rounds)
SELECT type, players, rounds FROM init
WHERE type = 2

Then mysql_insert_id() will return 33, not 36.
Wayne Theisinger
26-Feb-2004 07:40
In response to treylane at example dot com.

It is very very very important that you put in an "or die" or some other form of error handling.

Some scripts can fail invisibly and insert invalid data throughout your whole database because of mysql_insert_id
inserting the last successful insertid rather than reporting that the last attempt failed.

example of an or die statement.

$result = mysql_query($sql)
or die("Invalid query: " . mysql_error());
$EventID = mysql_insert_id();
dtez
12-Feb-2004 08:36
any zerofills on your id get chopped off on this function because it returns an int.
sly at noiretblanc dot org
14-Nov-2003 07:29
To get the NEXT insert id use the mysql query SHOW TABLE STATUS LIKE 'tablename' and get the field auto_increment...
treylane at example dot com
25-Apr-2003 05:36
This might be obvious, but it tripped me up - be careful when using last_insert_id() with persistent connections - running last_insert_id() after a failed update/insert/etc will return the last insert id of the last successful update/insert made by that CONNECTION rather than 0 for the number of rows updated by the previous non-working query, and who knows what the last query run on that connection was.
jameszhou2001 at yahoo dot ca
17-Dec-2002 10:30
Just a reminder, mysql_insert_id() should be called after 'mysql_affected_rows()', but BEFORE 'mysql_query("COMMIT")'.
frumler at the-beach dot no_spam dot net
04-Aug-2001 02:17
If you want to use the ID that was generated for one table and insert it into a second table, you can use SQL statements like this:

INSERT INTO foo (auto,text)
   VALUES(NULL,'text');              # generate ID by inserting NULL
INSERT INTO foo2 (id,text)
   VALUES(LAST_INSERT_ID(),'text');  # use ID in second table

...found here:
http://www.mysql.com/doc/en/Getting_unique_ID.html

It works even without inserting the NULL value for some reason ;)
The following is great for monitoring:
   $new_id = mysql_insert_id();
   print "New id: $new_id\n";

Hope it helps you all, cheers.
vksgeneric at hotmail dot com
10-Dec-1999 09:14
You can't do an INSERT DELAYED and expect to get anything but zero, for it runs in a separate thread, and mysql_insert_id() is tied to the current thread.
Vlad