ExecuteMany performance is insanely bad

[diwu-sf]

I know that the documentation makes it clear that executemany is a naive for loop:

No optimizations of the query (like batching) will be performed. 

But it's 2025 and please have a more optimized executemany that just issues a single SQL statement using VALUES(...) or something so this is actually usable in a data pipeline. Otherwise, trying to use the databricks SQL connector to write any non-trivial dataframe to a delta table is pointless.

Thanks!

[diwu-sf]

For reference, the ODBC databricks driver with cursor.fast_executemany = True can do ~140 rows / s. Still not great, but better than what this SQL connector can do