MYSQL Bulk “INSERT … ON DUPLICATE KEY UPDATE …” via “LOAD DATA INFILE”

I would recommend using INSERT...ON DUPLICATE KEY UPDATE.

If you use INSERT IGNORE, then the row won't actually be inserted if it results in a duplicate key. But the statement won't generate an error. It generates a warning instead. These cases include:

• Inserting a duplicate key in columns with PRIMARY KEY or UNIQUE constraints.
• Inserting a NULL into a column with a NOT NULL constraint.
• Inserting a row to a partitioned table, but the values you insert don't map to a partition.

If you use REPLACE, MySQL actually does a DELETE followed by an INSERT internally, which has some unexpected side effects:

• A new auto-increment ID is allocated.
• Dependent rows with foreign keys may be deleted (if you use cascading foreign keys) or else prevent the REPLACE.
• Triggers that fire on DELETE are executed unnecessarily.
• Side effects are propagated to replicas too.

correction: both REPLACE and INSERT...ON DUPLICATE KEY UPDATE are non-standard, proprietary inventions specific to MySQL. ANSI SQL 2003 defines a MERGE statement that can solve the same need (and more), but MySQL does not support the MERGE statement.

A user tried to edit this post (the edit was rejected by moderators). The edit tried to add a claim that INSERT...ON DUPLICATE KEY UPDATE causes a new auto-increment id to be allocated. It's true that the new id is generated, but it is not used in the changed row.

See demonstration below, tested with Percona Server 5.5.28. The configuration variable innodb_autoinc_lock_mode=1 (the default):

mysql> create table foo (id serial primary key, u int, unique key (u));
mysql> insert into foo (u) values (10);
mysql> select * from foo;
+----+------+
| id | u    |
+----+------+
|  1 |   10 |
+----+------+

mysql> show create table foo\G
CREATE TABLE `foo` (
  `id` bigint(20) unsigned NOT NULL AUTO_INCREMENT,
  `u` int(11) DEFAULT NULL,
  PRIMARY KEY (`id`),
  UNIQUE KEY `u` (`u`)
) ENGINE=InnoDB AUTO_INCREMENT=2 DEFAULT CHARSET=latin1

mysql> insert into foo (u) values (10) on duplicate key update u = 20;
mysql> select * from foo;
+----+------+
| id | u    |
+----+------+
|  1 |   20 |
+----+------+

mysql> show create table foo\G
CREATE TABLE `foo` (
  `id` bigint(20) unsigned NOT NULL AUTO_INCREMENT,
  `u` int(11) DEFAULT NULL,
  PRIMARY KEY (`id`),
  UNIQUE KEY `u` (`u`)
) ENGINE=InnoDB AUTO_INCREMENT=3 DEFAULT CHARSET=latin1

The above demonstrates that the IODKU statement detects the duplicate, and invokes the update to change the value of u. Note the AUTO_INCREMENT=3 indicates an id was generated, but not used in the row.

Whereas REPLACE does delete the original row and inserts a new row, generating and storing a new auto-increment id:

mysql> select * from foo;
+----+------+
| id | u    |
+----+------+
|  1 |   20 |
+----+------+
mysql> replace into foo (u) values (20);
mysql> select * from foo;
+----+------+
| id | u    |
+----+------+
|  3 |   20 |
+----+------+

Several tips:

1. Put inserts/updates in a transaction.
2. For older versions of SQLite - Consider a less paranoid journal mode (pragma journal_mode). There is NORMAL, and then there is OFF, which can significantly increase insert speed if you're not too worried about the database possibly getting corrupted if the OS crashes. If your application crashes the data should be fine. Note that in newer versions, the OFF/MEMORY settings are not safe for application level crashes.
3. Playing with page sizes makes a difference as well (PRAGMA page_size). Having larger page sizes can make reads and writes go a bit faster as larger pages are held in memory. Note that more memory will be used for your database.
4. If you have indices, consider calling CREATE INDEX after doing all your inserts. This is significantly faster than creating the index and then doing your inserts.
5. You have to be quite careful if you have concurrent access to SQLite, as the whole database is locked when writes are done, and although multiple readers are possible, writes will be locked out. This has been improved somewhat with the addition of a WAL in newer SQLite versions.
6. Take advantage of saving space...smaller databases go faster. For instance, if you have key value pairs, try making the key an INTEGER PRIMARY KEY if possible, which will replace the implied unique row number column in the table.
7. If you are using multiple threads, you can try using the shared page cache, which will allow loaded pages to be shared between threads, which can avoid expensive I/O calls.
8. Don't use !feof(file)!

I've also asked similar questions here and here.