bulkWrite

bulkWrite is a database operation that enables executing multiple write operations in a single call to the database driver. It is designed to improve throughput and reduce network overhead when applying large numbers of writes.

bulkWrite accepts an array of write operations, often referred to as write models. Typical operation types include insertOne, updateOne, updateMany, deleteOne, deleteMany, and replaceOne. Each operation specifies the target document(s) and the changes to apply. The result reflects the outcome of the entire bulk request.

Execution can be configured as ordered or unordered. In ordered mode, operations are executed sequentially and the bulk stops at the first error. In unordered mode, the operations may be processed in parallel, and the bulk continues despite individual errors. The choice affects error handling and performance.

The result of a bulkWrite operation includes a summary such as insertedCount, matchedCount, modifiedCount, deletedCount, and upsertedIds. Some drivers also return a bulk write error structure, including writeErrors and writeConcernErrors, if issues occur. Upserts are reported via upsertedIds and upsertedCount.

Atomicity and transactions: a single bulkWrite is not automatically atomic across all included operations. To guarantee atomicity for a set of writes, a multi-document transaction should be used. Bulk writes are commonly employed for bulk data import, batch updates, or large-scale maintenance tasks where reducing round trips and latency is important.