I am inclined to fix this by creating a connection directly, not via the JDBC DriverManager. But that has the potential to break people who are relying on the DriverManager. If this would affect you, please speak up by adding comments to the JIRA case.
This particular case seems to be related to shading. There was a similar issue in Flink [1] whose cause is not clear. Julian [1] https://issues.apache.org/jira/browse/FLINK-4581 > On Sep 1, 2021, at 8:35 AM, Steven Talbot (Jira) <j...@apache.org> wrote: > > Steven Talbot created CALCITE-4760: > -------------------------------------- > > Summary: Should not need to initialize a jdbc:calcite driver to > make a RelBuilder > Key: CALCITE-4760 > URL: https://issues.apache.org/jira/browse/CALCITE-4760 > Project: Calcite > Issue Type: Improvement > Reporter: Steven Talbot > > > We just tracked down a nasty issue with Calcite being shaded as a dependency > in our code that could cause a stack like > > > {noformat} > No suitable driver found for jdbc:calcite: > org/apache/calcite/tools/Frameworks.java:184:in withPrepare > org/apache/calcite/tools/RelBuilder.java:225:in create > {noformat} > {{}} > > {{}} > > when calling RelBuilder.create. > > Our fault ultimately, but [~julianhyde] pointed out that Calcite really > shouldn't be trying to make a JDBC connection here. > > {{}} > > > > -- > This message was sent by Atlassian Jira > (v8.3.4#803005)
