When invoked from a client node, are Ignite Custom SQL Functions executed on the server node or on the client node?
I think they are always executed on the server node before returning the values to the client node (in the case where the query is being executed from an ignite client node against a distributed cache on the ignite server node).
Is the above understanding correct? If so, i understand this is invoked & works at row level and that means i can even look up another cache inside the function to generate even a 'compute column'. Is that okay to do?
Thanks
Your understanding is correct, the SQL functions are always executed on server nodes. But I would be careful about accessing other caches or records from inside the function or generating any values. The function should be used for stateless scenarios because it can be executed more than once for a single record if the cluster topology changes while the function's query is being executed.
Share more details about your use case and we might suggest more suitable Ignite APIs for your task.