For databases like Microsoft SQL Server and MySQL, conn:merge{} provides a bulk insert option, which can improve performance when inserting a large number of rows.
The bulk insert feature ignores key column definitions and always performs an INSERT using a bulk insert syntax.
conn:merge{ data=T, bulk_insert=true} -- Enables bulk insert
Bulk insert operations only perform inserts. If you need to update rows as well, you must handle this separately. If the bulk insert encounters a duplicate row based on key columns, an error will occur.