Someone just logged https://issues.apache.org/jira/browse/CALCITE-3933 <https://issues.apache.org/jira/browse/CALCITE-3933> "Incorrect SQL Emitted for Unicode for Several Dialects”.
Consider the literal ‘schön’ (German for ‘beautiful’). It seems that we generate SQL standard syntax, u&’sch\00f6n’, which works on PostgreSQL, but several databases (including MySQL, Redshift, BigQuery) do not support it. What should we be generating? Is there an easy solution that will work across many databases. I was surprised to see that there are no tests for generating unicode literals. That’s part of the problem, I think. Julian