Redshift specific dialect and/or multi-row insert

[ahamidi]

We've been using this JDBC connector against Redshift and have run into a number of shortcomings.

Specifically there are 2 main issues, the first is the string -> TEXT -> VARCHAR(256) mapping (which essentially renders the auto.create functionality useless) and the second is the incredibly slow writes resulting from inserts records one at a time in Redshift.

I put together a rough spike here to test out some potential improvements and the results are promising. You can find my PR here:
https://github.com/meroxa/aiven-kafka-connect-jdbc/pull/1

I'd love to collaborate on getting this work done and shipped if there's interest.

[willyborankin]

Hi Ali, yes we are interested in this functionality. Could you please open PR (not the draft one)?

[ahmeroxa]

@willyborankin I've opened the PR here #39

[ivanyu]

I made a separate ticket for supporting multi-row inserts: #40. Based on the discussion in #39, this now depends on it.