I am new to Flink and trying to create a prototype that can do a continuous query as well as on-demand ones. The first part already works (with pre-defined queries), but struggling with the second. There is the SQL CLI that can be used. However, we want to be able to do this from a web endpoint.
I have seen Flink Jupyter notebook integrations which tells me this should work some way from a web interface. However, not sure how to do this without submitting an entire JAR each time.
So far, my research has lead me broadly to these two resources:
There is an outstanding request to implement this. See https://cwiki.apache.org/confluence/display/FLINK/FLIP-24+-+SQL+Client and https://cwiki.apache.org/confluence/display/FLINK/FLIP-91%3A+Support+SQL+Client+Gateway
There is an open source JDBC driver (have not tried this yet) - https://github.com/ververica/flink-jdbc-driver - not sure if this is maintained and says "This project is at an early stage".
Any suggestion on what else we can explore?
It's currently not supported in Flink itself. There are SQL clients offered by Ververica Platform, which is also available as a community edition. See https://www.ververica.com/apache-flink-sql-on-ververica-platform