switch CSV writer to the unix dialect so that newline chars are always quoted

[rs-sathya]

We're seeing the following error for some data that's being attempted to be imported: psycopg2.errors.BadCopyFileFormat: unquoted newline found in data

This could be because of newlines in the source stream. I am testing and will verify the below suggested fix.

So far, my proposed solution is to use the unix dialect of CSV writer: https://docs.python.org/3/library/csv.html#csv.unix_dialect

I am working on a short PR and we can resolve this issue once the PR is merged.