(mongodb >=1.0.0)
The MongoDB\Driver\BulkWrite collects one or more write operations that should be sent to the server. After adding any number of insert, update, and delete operations, the collection may be executed via MongoDB\Driver\Manager::executeBulkWrite().
Write operations may either be ordered (default) or unordered. Ordered write operations are sent to the server, in the order provided, for serial execution. If a write fails, any remaining operations will be aborted. Unordered operations are sent to the server in an arbitrary order where they may be executed in parallel. Any errors that occur are reported after all operations have been attempted.
Example #1 Mixed write operations are grouped by type
Mixed write operations (i.e. inserts, updates, and deletes) will be assembled into typed write commands to be sent sequentially to the server.
<?php
$bulk = new MongoDB\Driver\BulkWrite(['ordered' => true]);
$bulk->insert(['_id' => 1, 'x' => 1]);
$bulk->insert(['_id' => 2, 'x' => 2]);
$bulk->update(['x' => 2], ['$set' => ['x' => 1]]);
$bulk->insert(['_id' => 3, 'x' => 3]);
$bulk->delete(['x' => 1]);
?>
Will result in four write commands (i.e. roundtrips) being executed. Since the operations are ordered, the third insertion cannot be sent until the preceding update is executed.
Example #2 Ordered write operations causing an error
<?php
$bulk = new MongoDB\Driver\BulkWrite(['ordered' => true]);
$bulk->delete([]);
$bulk->insert(['_id' => 1]);
$bulk->insert(['_id' => 2]);
$bulk->insert(['_id' => 3, 'hello' => 'world']);
$bulk->update(['_id' => 3], ['$set' => ['hello' => 'earth']]);
$bulk->insert(['_id' => 4, 'hello' => 'pluto']);
$bulk->update(['_id' => 4], ['$set' => ['hello' => 'moon']]);
$bulk->insert(['_id' => 3]);
$bulk->insert(['_id' => 4]);
$bulk->insert(['_id' => 5]);
$manager = new MongoDB\Driver\Manager('mongodb://localhost:27017');
$writeConcern = new MongoDB\Driver\WriteConcern(MongoDB\Driver\WriteConcern::MAJORITY, 1000);
try {
$result = $manager->executeBulkWrite('db.collection', $bulk, $writeConcern);
} catch (MongoDB\Driver\Exception\BulkWriteException $e) {
$result = $e->getWriteResult();
// Check if the write concern could not be fulfilled
if ($writeConcernError = $result->getWriteConcernError()) {
printf("%s (%d): %s\n",
$writeConcernError->getMessage(),
$writeConcernError->getCode(),
var_export($writeConcernError->getInfo(), true)
);
}
// Check if any write operations did not complete at all
foreach ($result->getWriteErrors() as $writeError) {
printf("Operation#%d: %s (%d)\n",
$writeError->getIndex(),
$writeError->getMessage(),
$writeError->getCode()
);
}
} catch (MongoDB\Driver\Exception\Exception $e) {
printf("Other error: %s\n", $e->getMessage());
exit;
}
printf("Inserted %d document(s)\n", $result->getInsertedCount());
printf("Updated %d document(s)\n", $result->getModifiedCount());
?>
以上例程会输出：
Operation#7: E11000 duplicate key error index: db.collection.$_id_ dup key: { : 3 } (11000)
Inserted 4 document(s)
Updated 2 document(s)
If the write concern could not be fullfilled, the example above would output something like:
waiting for replication timed out (64): array (
'wtimeout' => true,
)
Operation#7: E11000 duplicate key error index: databaseName.collectionName.$_id_ dup key: { : 3 } (11000)
Inserted 4 document(s)
Updated 2 document(s)
If we execute the example above, but allow for unordered writes:
<?php
$bulk = new MongoDB\Driver\BulkWrite(['ordered' => false]);
/* ... */
?>
以上例程会输出：
Operation#7: E11000 duplicate key error index: db.collection.$_id_ dup key: { : 3 } (11000)
Operation#8: E11000 duplicate key error index: db.collection.$_id_ dup key: { : 4 } (11000)
Inserted 5 document(s)
Updated 2 document(s)